Advertisement

Patterns and Risk Factor Profiles of Visual Loss in a Multiethnic Asian Population: The Singapore Epidemiology of Eye Diseases Study

  • Tien Y. Wong
    Correspondence
    Inquiries to Tien Y. Wong, Singapore Eye Research Institute, Singapore National Eye Center, 11 Third Hospital Avenue, Singapore 168751
    Affiliations
    Singapore Eye Research Institute, Singapore National Eye Centre, Singapore

    Duke-NUS Medical School, Singapore

    Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore
    Search for articles by this author
  • Yih-Chung Tham
    Affiliations
    Singapore Eye Research Institute, Singapore National Eye Centre, Singapore

    Duke-NUS Medical School, Singapore
    Search for articles by this author
  • Charumathi Sabanayagam
    Affiliations
    Singapore Eye Research Institute, Singapore National Eye Centre, Singapore

    Duke-NUS Medical School, Singapore
    Search for articles by this author
  • Ching-Yu Cheng
    Affiliations
    Singapore Eye Research Institute, Singapore National Eye Centre, Singapore

    Duke-NUS Medical School, Singapore

    Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore
    Search for articles by this author

      Purpose

      To describe the pattern and risk factor traits associated with visual loss (defined as either visual impairment [VI] or blindness) in a multiethnic Asian cohort of Malay, Indian, and Chinese individuals living in Singapore.

      Methods

      A total of 10 020 participants from the Singapore Epidemiology of Eye Diseases Study were examined between 2004 and 2011. All underwent standardized examinations. VI (visual acuity <20/40 to ≥20/200) and blindness (visual acuity <20/200) were defined based on the US definition, better-seeing eye. Singapore Population Census 2010 was used to calculate age-standardized prevalence. Multiple logistic regression analysis was performed to determine the independent and joint risk factors associated with visual loss.

      Results

      Malay individuals had higher age-standardized prevalence of best-corrected and presenting VI (5.4% and 19.9%, respectively) than Indian (3.6% and 18.0%) and Chinese individuals (3.3% and 17.2%). Cataract was the main cause for presenting and best-corrected blindness; cataract and diabetic retinopathy were the top causes for best-corrected VI, consistently observed across the 3 ethnic groups. Older age, female sex, lower socioeconomic status, diabetes, systemic comorbidities, and cognitive impairment were independently associated with increased risk of best-corrected visual loss (all P ≤ .027). Individuals aged ≥60 years with diabetes were 12.7 times (95% confidence interval, 8.39–19.23) likely to have best-corrected visual loss, compared with younger, nondiabetic individuals. Lower income and education explained 58.1% and 23.2% of best-corrected visual loss in this population, respectively.

      Conclusion

      In this urban multiethnic Asian population, we identified common traits associated with visual loss across Malay, Indian, and Chinese individuals. These results will be useful for the planning and designing of eye health services and strategies for Asia's rapidly developing populations living in urban communities. NOTE: Publication of this article is sponsored by the American Ophthalmological Society.
      Visual loss, defined as either visual impairment (VI) or blindness, is a major public health problem,
      • Bourne R.R.A.
      • Flaxman S.R.
      • Braithwaite T.
      • et al.
      Magnitude, temporal trends, and projections of the global prevalence of blindness and distance and near vision impairment: a systematic review and meta-analysis.
      and is among the top 3 most common impairments in terms of years lived with disability.
      Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: a systematic analysis for the Global Burden of Disease Study 2015.
      Visual loss is associated with reduced quality of life, increased risk of frailty, and mortality.
      • Lamoureux E.L.
      • Chong E.W.
      • Thumboo J.
      • et al.
      Vision impairment, ocular conditions, and vision-specific function: the Singapore Malay Eye Study.
      • Siantar R.G.
      • Cheng C.Y.
      • Gemmy Cheung C.M.
      • et al.
      Impact of visual impairment and eye diseases on mortality: the Singapore Malay Eye Study (SiMES).
      • Dai W.
      • Tham Y.C.
      • Chee M.L.
      • et al.
      Falls and recurrent falls among adults in a multi-ethnic Asian population: The Singapore Epidemiology of Eye Diseases Study.
      Globally, it was estimated that 36 million people were blind and 400 million suffered from VI in 2015.
      • Bourne R.R.A.
      • Flaxman S.R.
      • Braithwaite T.
      • et al.
      Magnitude, temporal trends, and projections of the global prevalence of blindness and distance and near vision impairment: a systematic review and meta-analysis.
      Asia alone accounts for approximately 60% cases of VI and blindness worldwide.
      • Bourne R.R.A.
      • Flaxman S.R.
      • Braithwaite T.
      • et al.
      Magnitude, temporal trends, and projections of the global prevalence of blindness and distance and near vision impairment: a systematic review and meta-analysis.
      • Pascolini D.
      • Mariotti S.P.
      Global estimates of visual impairment: 2010.
      Furthermore, Asia accounts for most of the most common eye diseases; for example, 35% (59 million) of age-related macular degeneration (AMD)
      • Wong W.L.
      • Su X.
      • Li X.
      • et al.
      Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis.
      and 60% (39 million) of glaucoma cases globally.
      • Tham Y.C.
      • Li X.
      • Wong T.Y.
      • Quigley H.A.
      • Aung T.
      • Cheng C.Y.
      Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis.
      Taken together, there is an increase and possibly disproportionate burden of visual loss and age-related eye diseases on Asian individuals. This burden is likely to further increase in the next decade, given the rapid aging trends in Asia.
      United Nations Department of Economic and Social Affairs
      World Population Prospects: The 2017 Revision.
      • Bureau P.R.
      World Population Data Sheet.
      Hence, there is a need to have a granular understanding of the patterns of VI and blindness among Asian individuals, and the major factors associated with visual loss.
      In the past 2 decades, there have been multiple population-based epidemiological studies and surveys in Asia that have provided considerable information on the patterns and trends of VI and blindness in disparate countries from Cambodia,
      • Rutzen A.R.
      • Ellish N.J.
      • Schwab L.
      • et al.
      Blindness and eye disease in Cambodia.
      China,
      • Zhao J.
      • Jia L.
      • Sui R.
      • Ellwein L.B.
      Prevalence of blindness and cataract surgery in Shunyi County, China.
      • Xu L.
      • Wang Y.
      • Li Y.
      • et al.
      Causes of blindness and visual impairment in urban and rural areas in Beijing: the Beijing Eye Study.
      • Huang S.
      • Zheng Y.
      • Foster P.J.
      • Huang W.
      • He M.
      Prevalence and causes of visual impairment in Chinese adults in urban southern China.
      • Zhao J.
      • Ellwein L.B.
      • Cui H.
      • et al.
      Prevalence of vision impairment in older adults in rural China: the China Nine-Province Survey.
      • Liang Y.B.
      • Friedman D.S.
      • Wong T.Y.
      • et al.
      Prevalence and causes of low vision and blindness in a rural chinese adult population: the Handan Eye Study.
      • Li Z.
      • Cui H.
      • Liu P.
      • Zhang L.
      • Yang H.
      • Zhang L.
      Prevalence and causes of blindness and visual impairment among the elderly in rural southern Harbin, China.
      • Li X.
      • Zhou Q.
      • Sun L.
      • et al.
      Prevalence of blindness and low vision in a rural population in northern China: preliminary results from a population-based survey.
      • Li J.
      • Zhong H.
      • Cai N.
      • et al.
      The prevalence and causes of visual impairment in an elderly Chinese Bai ethnic rural population: the Yunnan minority eye study.
      • Tang Y.
      • Wang X.
      • Wang J.
      • et al.
      Prevalence and causes of visual impairment in a Chinese adult population: the Taizhou Eye Study.
      • Yang W.Y.
      • Li J.
      • Zhao C.H.
      • et al.
      Population-based assessment of visual impairment among ethnic Dai adults in a rural community in China.
      • Zhao J.
      • Xu X.
      • Ellwein L.B.
      • et al.
      Prevalence of vision impairment in older adults in rural China in 2014 and comparisons with the 2006 China nine-province survey.
      India,
      • Thulasiraj R.D.
      • Nirmalan P.K.
      • Ramakrishnan R.
      • et al.
      Blindness and vision impairment in a rural south Indian population: the Aravind Comprehensive Eye Survey.
      • Murthy G.V.
      • Gupta S.
      • Ellwein L.B.
      • Munoz S.R.
      • Bachani D.
      • Dada V.K.
      A population-based eye survey of older adults in a rural district of Rajasthan: I. Central vision impairment, blindness, and cataract surgery.
      • Thulasiraj R.D.
      • Rahamathulla R.
      • Saraswati A.
      • Selvaraj S.
      • Ellwein L.B.
      The Sivaganga eye survey: I. Blindness and cataract surgery.
      • Dandona R.
      • Dandona L.
      • Srinivas M.
      • et al.
      Moderate visual impairment in India: the Andhra Pradesh Eye Disease Study.
      • Nirmalan P.K.
      • Thulasiraj R.D.
      • Maneksha V.
      • et al.
      A population based eye survey of older adults in Tirunelveli district of south India: blindness, cataract surgery, and visual outcomes.
      • Vijaya L.
      • George R.
      • Arvind H.
      • et al.
      Prevalence and causes of blindness in the rural population of the Chennai Glaucoma Study.
      • Nangia V.
      • Jonas J.B.
      • Gupta R.
      • Khare A.
      • Sinha A.
      Visual impairment and blindness in rural central India: the Central India Eye and Medical Study.
      Japan,
      • Iwase A.
      • Araie M.
      • Tomidokoro A.
      • Yamamoto T.
      • Shimizu H.
      • Kitazawa Y.
      Prevalence and causes of low vision and blindness in a Japanese adult population: the Tajimi Study.
      • Nakamura Y.
      • Tomidokoro A.
      • Sawaguchi S.
      • Sakai H.
      • Iwase A.
      • Araie M.
      Prevalence and causes of low vision and blindness in a rural Southwest Island of Japan: the Kumejima study.
      Korea,
      • Rim T.H.
      • Nam J.S.
      • Choi M.
      • Lee S.C.
      • Lee C.S.
      Prevalence and risk factors of visual impairment and blindness in Korea: the Fourth Korea National Health and Nutrition Examination Survey in 2008-2010.
      Nepal,
      • Thapa S.S.
      • Berg R.V.
      • Khanal S.
      • et al.
      Prevalence of visual impairment, cataract surgery and awareness of cataract and glaucoma in Bhaktapur district of Nepal: the Bhaktapur Glaucoma Study.
      Singapore,
      • Saw S.M.
      • Foster P.J.
      • Gazzard G.
      • Seah S.
      Causes of blindness, low vision, and questionnaire-assessed poor visual function in Singaporean Chinese adults: The Tanjong Pagar Survey.
      Pakistan,
      • Ahmad K.
      • Khan M.D.
      • Qureshi M.B.
      • et al.
      Prevalence and causes of blindness and low vision in a rural setting in Pakistan.
      Sri Lanka,
      • Edussuriya K.
      • Sennanayake S.
      • Senaratne T.
      • et al.
      The prevalence and causes of visual impairment in central Sri Lanka the Kandy Eye study.
      and Taiwan
      • Hsu W.M.
      • Cheng C.Y.
      • Liu J.H.
      • Tsai S.Y.
      • Chou P.
      Prevalence and causes of visual impairment in an elderly Chinese population in Taiwan: the Shihpai Eye Study.
      (Supplementary Table 1). However, 3 major gaps remain.

      Lack of Evaluation on Ethnic Differences Among Asian Individuals

      The influence of ethnicity on health is myriad, affecting not only the incidence, severity, and prognosis of diseases but also the utilization of health services.
      • Anderson J.J.
      • Felson D.T.
      Factors associated with osteoarthritis of the knee in the first national Health and Nutrition Examination Survey (HANES I). Evidence for an association with overweight, race, and physical demands of work.
      • Johnston J.
      Improving utilisation of dental services by understanding cultural difference.
      • Otten M.W.
      • Teutsch S.M.
      • Williamson D.F.
      • Marks J.S.
      The effect of known risk factors on the excess mortality of black adults in the United States.
      Ethnicity affects health through complex gene-environment and behavior-environment interactions, and research on multiethnic populations has been advocated by renowned health organizations such as the American Heart Association and the National Health Service Scotland.
      • Goff D.C.
      • Lloyd-Jones D.M.
      • Bennett G.
      • et al.
      2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines.

      Edinburgh: NHS Health Scotland SG, 2009,1–84. The Scottish Ethnicity and Health Research Strategy Working Group. Health in our multi-ethnic Scotland future research priorities (full report). 2009. Available at http://www.healthscotland.com/documents/5720.aspx. Accessed July 10, 2019.

      In Singapore, previous studies also showed the importance of evaluating ethnic differences when studying systemic diseases. For example, Singapore's National Health Survey previously observed Indian and Malay individuals to have higher prevalence of diabetes compared with Chinese,
      • Tan C.E.
      • Emmanuel S.C.
      • Tan B.Y.
      • Jacob E.
      Prevalence of diabetes and ethnic differences in cardiovascular risk factors. The 1992 Singapore National Health Survey.
      thus leading to further postulation that prevalence of diabetic-related complications, such as diabetic retinopathy (DR), also may differ across ethnic groups in Singapore. Furthermore, a previous multiethnic coronary heart disease study on diabetic Singaporeans also reported ethnicity to be independently associated with coronary heart disease–related morbidity.
      • Yeo K.
      • Tai B.
      • Heng D.
      • et al.
      Ethnicity modifies the association between diabetes mellitus and ischaemic heart disease in Chinese, Malays and Asian Indians living in Singapore.
      Collectively, these examples further emphasize the relevance of evaluating ethnic comparison in a country with multiethnic composition such as Singapore.
      In the aspect of VI and blindness, previous multiethnic studies in the United States provided important insights into the understanding of the relative burden, risk factors, and impact of visual loss between white and African American individuals.
      • Fisher D.E.
      • Shrager S.
      • Shea S.J.
      • et al.
      Visual impairment in White, Chinese, Black, and Hispanic participants from the Multi-Ethnic Study of Atherosclerosis Cohort.
      • Tielsch J.M.
      • Sommer A.
      • Witt K.
      • Katz J.
      • Royall R.M.
      Blindness and visual impairment in an American urban population. The Baltimore Eye Survey.
      • Klein R.
      • Klein B.E.
      The prevalence of age-related eye diseases and visual impairment in aging: current estimates.
      Findings from the US multiethnic studies rendered an extrapolated basis to hypothesize that ethnic variations also may be observed among Asian individuals as well, especially given the heterogeneity across different Asian ethnic groups.
      However, currently, there is a palpable lack of evaluation on ethnic differences among Asian individuals on the risk of VI and blindness. Most previous studies have mainly focused on evaluations of specific individual ethnic groups (eg, Chinese, Indian), and few have evaluated possible ethnic difference among Asian individuals. This is mainly because the sampled population was typically more homogeneous by ethnicity. Thus, direct comparisons of ethnic variations (eg, Chinese vs Indian) between these previous Asian studies cannot be made accurately because of the inherent variations in population characteristics across these studies. Hence, ethnic comparison among Asian individuals remains an unaddressed knowledge gap.
      In this regard, Singapore offers a unique opportunity to study variations among the 3 main Asian ethnic groups: Chinese, Indian, and Malay. Previous work had also observed differences in genetic ancestral profile across these 3 groups,
      • Tham Y.C.
      • Liao J.
      • Vithana E.N.
      • et al.
      Aggregate effects of intraocular pressure and cup-to-disc ratio genetic variants on glaucoma in a multiethnic Asian Population.
      further justifying the rationale for ethnic comparison. On the other hand, genetic ancestral analyses showed that Singaporean Indian individuals well resemble Southern Indian,
      • Wu D.
      • Dou J.
      • Chai X.
      • et al.
      Large-scale whole-genome sequencing of three diverse Asian populations in Singapore.
      Singaporean Chinese resemble Southern Han Chinese,
      • Wu D.
      • Dou J.
      • Chai X.
      • et al.
      Large-scale whole-genome sequencing of three diverse Asian populations in Singapore.
      and Singaporean Malay individuals genetically resemble Peninsular Malaysian Malay and Indonesian Malay individuals.
      • Deng L.
      • Hoh B.P.
      • Lu D.
      • et al.
      Dissecting the genetic structure and admixture of four geographical Malay populations.
      Taken together, this indicates that evaluation of these 3 ethnic groups in Singapore also may provide useful proxy information on Malay, Indian, and Chinese individuals residing in other Asian countries.
      Interesting previous observations on inherent differences among the 3 ethnic groups include the following examples. The eyes of Chinese individuals have been observed to have longer axial length and are more likely to be myopic,
      • Tham Y.C.
      • Aung T.
      • Fan Q.
      • et al.
      Joint Effects of Intraocular Pressure and Myopia on Risk of Primary Open-Angle Glaucoma: The Singapore Epidemiology of Eye Diseases Study.
      • Chua J.
      • Tham Y.C.
      • Liao J.
      • et al.
      Ethnic differences of intraocular pressure and central corneal thickness: the Singapore Epidemiology of Eye Diseases study.
      Indian individuals have thinner cornea and higher intraocular pressure,
      • Chua J.
      • Tham Y.C.
      • Liao J.
      • et al.
      Ethnic differences of intraocular pressure and central corneal thickness: the Singapore Epidemiology of Eye Diseases study.
      Malay individuals are more likely to be hypertensive,
      • Chua J.
      • Tham Y.C.
      • Liao J.
      • et al.
      Ethnic differences of intraocular pressure and central corneal thickness: the Singapore Epidemiology of Eye Diseases study.
      and Indian individuals are more likely to have diabetes.
      • Tan G.S.
      • Gan A.
      • Sabanayagam C.
      • et al.
      Ethnic differences in the prevalence and risk factors of diabetic retinopathy: the Singapore Epidemiology of Eye Diseases Study.
      Furthermore, Chinese have a higher prevalence of AMD
      • Chua J.
      • Tham Y.C.
      • Liao J.
      • et al.
      Ethnic differences of intraocular pressure and central corneal thickness: the Singapore Epidemiology of Eye Diseases study.
      and primary angle closure glaucoma.
      • Tham Y.C.
      • Li X.
      • Wong T.Y.
      • Quigley H.A.
      • Aung T.
      • Cheng C.Y.
      Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis.
      • Foster P.J.
      • Oen F.T.
      • Machin D.
      • et al.
      The prevalence of glaucoma in Chinese residents of Singapore: a cross-sectional population survey of the Tanjong Pagar district.
      • Shen S.Y.
      • Wong T.Y.
      • Foster P.J.
      • et al.
      The prevalence and types of glaucoma in malay people: the Singapore Malay eye study.
      • Narayanaswamy A.
      • Baskaran M.
      • Zheng Y.
      • et al.
      The prevalence and types of glaucoma in an urban Indian population: the Singapore Indian Eye Study.
      Indian individuals have higher prevalence of DR compared with Malay and Chinese individual.
      • Tan G.S.
      • Gan A.
      • Sabanayagam C.
      • et al.
      Ethnic differences in the prevalence and risk factors of diabetic retinopathy: the Singapore Epidemiology of Eye Diseases Study.
      Consistent with these observations, early registry studies in Singapore also have shown ethnicity as an important risk factor for acute angle closure glaucoma, cataract surgery, and retinal detachment.
      • Wong T.Y.
      • Tielsch J.M.
      • Schein O.D.
      Racial difference in the incidence of retinal detachment in Singapore.
      • Wong T.Y.
      • Foster P.J.
      • Seah S.K.
      • Chew P.T.
      Rates of hospital admissions for primary angle closure glaucoma among Chinese, Malays, and Indians in Singapore.
      • Wong T.Y.
      Cataract extraction rates among Chinese, Malays, and Indians in Singapore: a population-based analysis.
      Considering the heterogeneity in genetic, ocular, and systemic profiles across the 3 ethnicities, it is therefore plausible to hypothesize that the patterns of VI and blindness may naturally differ across ethnic groups in Asia. Taken together, this further highlights the need for comprehensive evaluation in this aspect. As Malay, Indian, and Chinese individuals account for approximately 45% of the world population, with a global estimate of approximately 3.5 billion persons,
      United Nations Department of Economic and Social Affairs
      World Population Prospects: The 2017 Revision.
      data on VI and blindness on these 3 main representative ethnic groups in Asia is crucial.

      Lack of Asian Urban Studies on VI and Blindness

      A second knowledge gap exists with the rapid urbanization of Asian populations. Most previous Asian studies in the past 2 decades have understandably been focused on rural regions and communities (Supplementary Table 1). In fact, only a handful of studies from Korea, Japan, and Taiwan
      • Iwase A.
      • Araie M.
      • Tomidokoro A.
      • Yamamoto T.
      • Shimizu H.
      • Kitazawa Y.
      Prevalence and causes of low vision and blindness in a Japanese adult population: the Tajimi Study.
      • Nakamura Y.
      • Tomidokoro A.
      • Sawaguchi S.
      • Sakai H.
      • Iwase A.
      • Araie M.
      Prevalence and causes of low vision and blindness in a rural Southwest Island of Japan: the Kumejima study.
      • Rim T.H.
      • Nam J.S.
      • Choi M.
      • Lee S.C.
      • Lee C.S.
      Prevalence and risk factors of visual impairment and blindness in Korea: the Fourth Korea National Health and Nutrition Examination Survey in 2008-2010.
      • Hsu W.M.
      • Cheng C.Y.
      • Liu J.H.
      • Tsai S.Y.
      • Chou P.
      Prevalence and causes of visual impairment in an elderly Chinese population in Taiwan: the Shihpai Eye Study.
      have evaluated the trends of visual loss in urban regions. Hence, there remains currently a significant underrepresentation of Asian studies from urban communities. This is especially relevant with the current rapid development and urbanization in Asian countries. Furthermore, given that the rate of aging is more marked in urban regions compared with rural regions in Asia,
      United Nations Department of Economic and Social Affairs
      World Population Prospects: The 2017 Revision.
      and the concomitant increases in the prevalence of age-related systemic chronic diseases associated with urbanization (eg, hypertension, diabetes),
      • Gong P.
      • Liang S.
      • Carlton E.J.
      • et al.
      Urbanisation and health in China.
      • Strong K.
      • Mathers C.
      • Leeder S.
      • Beaglehole R.
      Preventing chronic diseases: how many lives can we save?.
      the number of elderly in urban communities is expected to further increase and to be accompanied with corresponding increase in VI cases.
      Because of the differences in environmental exposures, accessibility to eye care, and lifestyle profiles between urban and rural communities, it is likely that findings from urban study may offer new insights that are different from previous rural reports. Furthermore, within urban populations, there is generally a “wider socioeconomic spectrum” and thus greater socioeconomic inequality and disparity compared with rural populations. As lower socioeconomic status had been shown to be associated with poorer eye care accessibility and utility,
      • Ehrlich J.R.
      • Stagg B.C.
      • Andrews C.
      • Kumagai A.
      • Musch D.C.
      Vision impairment and receipt of eye care among older adults in low- and middle-income countries.
      evaluation in an urban setting provides a good opportunity to further identify socioeconomic factors associated with visual loss, further providing concrete evidence in this aspect to better inform policy makers in designing relevant social plans and policies.

      Current Knowledge Gaps on Risk Factors Associated With VI and Blindness

      Finally, previous population-based studies mainly focused on reporting the prevalence and causes of visual loss, with very few that have evaluated comprehensively the entire domains of demographic, systemic, socioeconomic, and lifestyle profiles associated with VI and blindness. In fact, even risk factor evaluation on visual loss also has not been extensively reported in Western landmark population-based studies, such as Beaver Dam Eye Study,
      • Klein R.
      • Klein B.E.
      • Linton K.L.
      • De Mets D.L.
      The Beaver Dam Eye Study: visual acuity.
      Blue Mountains Eye Study,
      • Attebo K.
      • Mitchell P.
      • Smith W.
      Visual acuity and the causes of visual loss in Australia. The Blue Mountains Eye Study.
      and the Rotterdam Eye Study.
      • Varma R.
      • Ying-Lai M.
      • Klein R.
      • Azen S.P.
      Los Angeles Latino Eye Study Group
      Prevalence and risk indicators of visual impairment and blindness in Latinos: the Los Angeles Latino Eye Study.
      This is partly because of the limited/incomplete capture of relevant systemic or socioeconomic data or insufficient sample size in previous studies. Altogether, this presents as a missed opportunity, as insights into the characteristics associated with visual loss will potentially aid in formulation of more effective interventional strategies to combat VI and blindness.
      Furthermore, although a limited number of studies have evaluated the independent effects of demographic and other factors on VI and blindness,
      • Xu L.
      • Wang Y.
      • Li Y.
      • et al.
      Causes of blindness and visual impairment in urban and rural areas in Beijing: the Beijing Eye Study.
      • Zhao J.
      • Xu X.
      • Ellwein L.B.
      • et al.
      Prevalence of vision impairment in older adults in rural China in 2014 and comparisons with the 2006 China nine-province survey.
      most previous analyses were brief and none have comprehensively explored joint and combined effects of various demographic, systemic, socioeconomic, and lifestyle factors with VI and blindness. Questions that could be addressed include the following: are older people with diabetes and hypertension at higher risk of visual loss than younger people without systemic comorbidities? Joint effect analysis provides more holistic and detailed risk factor profiling; such information will be informative in fine-tuning risk stratification for visual loss.

      The Singapore Epidemiology of Eye Diseases Study

      The Singapore Epidemiology of Eye Disease (SEED) study is a world-leading epidemiology research program, commenced in 2004. The overall aim of the SEED study is to establish a “data portal” and information source on epidemiology of Asian eye diseases that covers the entire spectrum of prevalence, incidence, risk factors, and impact of major Asian eye diseases, based on the 3 major ethnic groups (Malay, Indian, Chinese) in Singapore. Evaluation of this multiethnic Asian population as a whole and ethnic differences among the ethnic groups, are 2 pivot points for SEED. Singapore is the ideal “population laboratory” for this purpose because of its geographic compactness and the inclusive reach of the local government's policies and urbanization. The SEED study was modeled after earlier landmark population-based eye studies with strong scientific reputations, namely the Beaver Dam Eye Study (United States, commenced in 1987), and the Blue Mountains Eye Study (Australia, commenced in 1992). Standardized protocols from SEED were modified from these 2 predecessors; this created a unique opportunity to articulate ocular research findings from Singapore among the global elites of population-based eye studies.
      The SEED study is one of the few very well characterized cohorts internationally with a rich reservoir of data that is not only limited to eye diseases, but also encompasses systemic disease data, bio-samples, high-quality digital imaging data, data on quality of life assessment, and genetic markers. To date, the SEED program has resulted in many novel findings and generated close to 350 publications. Data collected from the SEED study also have been used by national and international agencies (eg, the Ministry of Health of Singapore, World Health Organization, the Global Burden of Disease program
      • Bourne R.R.
      • Stevens G.A.
      • White R.A.
      • et al.
      Causes of vision loss worldwide, 1990-2010: a systematic analysis.
      • Flaxman S.R.
      • Bourne R.R.A.
      • Resnikoff S.
      • et al.
      Global causes of blindness and distance vision impairment 1990-2020: a systematic review and meta-analysis.
      ), health bodies to formulate clinical guidelines (eg, 2014 Ministry of Health Diabetes Guidelines,
      • Goh S.Y.
      • Ang S.B.
      • Bee Y.M.
      • et al.
      Ministry of Health clinical practice guidelines: diabetes mellitus.
      2016 Asia Pacific Glaucoma Guidelines, 2016 American Diabetes Association Guidelines,
      • Sabanayagam C.
      • Khoo E.Y.
      • Lye W.K.
      • et al.
      Diagnosis of diabetes mellitus using HbA1c in Asians: relationship between HbA1c and retinopathy in a multiethnic Asian population.
      2017 International Council of Ophthalmology Diabetic Eye Care Guidelines
      • Wong T.Y.
      • Aiello L.P.
      • Ferris F.
      • et al.
      ICO Guidelines for Diabetic Eye Care (Updated 2017).
      ), and have provided estimates of eye disease burden to set up Singapore's national DR screening,
      • Nguyen H.V.
      • Tan G.S.
      • Tapp R.J.
      • et al.
      Cost-effectiveness of a national telemedicine diabetic retinopathy screening program in Singapore.
      and to assist the local government's planning for future health care manpower (future ophthalmology and optometry manpower
      • Ansah J.P.
      • De Korne D.
      • Bayer S.
      • et al.
      Future requirements for and supply of ophthalmologists for an aging population in Singapore.
      ) in Singapore.
      The SEED study is currently the single largest population-based comprehensive eye study comprising 10 033 adults aged 40 years and older, composed of Malay, Indian, and Chinese individuals, the 3 major ethnic groups in Asia. Standardized setting and examinations were performed across the 3 ethnic groups, thus providing a unique and unprecedented opportunity to compare data across these 3 ethnic groups in a common geographic and socioeconomic environment. To date, this has not been done in any other population-based studies worldwide. The unique design of the SEED study in an urban setting, coupled with a large sample and wide range of well-characterized data, will help to adequately address the previously mentioned knowledge gaps, and may provide new insights into the trends and patterns of racial differences in VI and blindness among Asian individuals.
      Previously, we had only published prevalence of VI and blindness, separately for Malay, Indian, and Chinese individuals.
      • Zheng Y.
      • Lavanya R.
      • Wu R.
      • et al.
      Prevalence and causes of visual impairment and blindness in an urban Indian population: the Singapore Indian Eye Study.
      • Tham Y.C.
      • Lim S.H.
      • Shi Y.
      • et al.
      Trends of visual impairment and blindness in the Singapore Chinese population over a decade.
      • Wong T.Y.
      • Chong E.W.
      • Wong W.L.
      • et al.
      Prevalence and causes of low vision and blindness in an urban malay population: the Singapore Malay Eye Study.
      In this current thesis, we highlight areas that were not published/done previously, and more important, new areas that can now be explored in this thesis, given the large, unprecedented combined data of the 3 ethnic groups. First, we previously did not evaluate the overall trends and ethnic difference of VI and blindness across the 3 major Asian ethnic groups in Singapore. To date, this has not been done in any other population-based studies as well. Second, we previously also did not evaluate/publish systemic and socioeconomic factors that are independently associated with visual loss. With the broad range of systemic and socioeconomic data captured in our SEED study, we are poised to provide new insights in these aspects. Third, presence of multiple comorbidity of systemic diseases is common in aging populations. Nevertheless, evaluation in this area is often limited by power/sample size. Hence, the potential impact of multiple comorbidity on visual loss has not been reported in other studies. By combining the 3 population-based data from SEED, we now have better statistical power to comprehensively evaluate the combined effects of ethnicity, socioeconomic risk factors, and multimorbidity of systemic diseases on visual loss. Findings in this area are particularly relevant for aging populations in urban communities.

       Central Hypotheses

      Based on the previously mentioned knowledge gaps and rationale, a few hypotheses were formulated. First, given the differences in lifestyle, socioeconomic profiles, and cultural facets across the 3 ethnic groups, coupled with observations from previous reports that indicated differences in health awareness and health-seeking behavior across ethnic groups,
      • Lim R.B.
      • Zheng H.
      • Yang Q.
      • Cook A.R.
      • Chia K.S.
      • Lim W.Y.
      Ethnic and gender specific life expectancies of the Singapore population, 1965 to 2009 - converging, or diverging?.
      • Chan C.Q.H.
      • Lee K.H.
      • Low L.L.
      A systematic review of health status, health seeking behaviour and healthcare utilisation of low socioeconomic status populations in urban Singapore.
      we hypothesized that Malay individuals have a higher prevalence of VI and blindness, whereas Chinese have the lowest prevalence among the 3 groups. Second, we hypothesized that lower socioeconomic indicators and presence of systemic diseases, such as diabetes, are associated with higher odds of prevalent VI and blindness. Third, we postulated that presence of combined measures of ethnicity, socioeconomic risk factors, and multiple comorbidity of systemic diseases increases the odds of VI and blindness.

       Objective of Thesis and Potential Impact

      The objective of this thesis was twofold. First, we attempted to describe and examine ethnic differences in patterns and causes of VI and blindness in a multiethnic Asian population of Malay, Indian, and Chinese individuals living in an Asian urban community. Second, we sought to assess the independent and joint effects of demographic, systemic, socioeconomic, and lifestyle factors on VI and blindness in this multiethnic Asian population.
      This thesis provides the first large population-based summary of pattern, trends, and factors associated with VI and blindness in a multiethnic Asian cohort that represents the 3 major ethnic groups in Asia. Findings from this thesis will be particularly relevant to the increasing number of individuals with visual loss in Asia due to rapid aging.
      • Bourne R.R.A.
      • Flaxman S.R.
      • Braithwaite T.
      • et al.
      Magnitude, temporal trends, and projections of the global prevalence of blindness and distance and near vision impairment: a systematic review and meta-analysis.
      In addition, independent and joint effect analysis of risk factors associated with visual loss may provide new knowledge to better stratify and identify individuals who are at higher risk of having visual loss, thus aiding public health officials in formulating more targeted vision-screening strategies. Overall, findings from this thesis will be useful and relevant in the planning and designing of eye health services for Asia's rapidly developing urban communities. Collectively, these data will directly and indirectly contribute to the World Health Organization (WHO)'s VISION 2020 goal of reducing VI and blindness.
      • World Health Organization
      Universal Eye Health: A Global Action Plan, 2014-2019.

      Methods

       Study Design

      The SEED study is a population-based cross-sectional study, comprising 3 major ethnic groups in Singapore: Malay (the Singapore Malay Eye Study, year 2004 to 2006), Indian (the Singapore Indian Eye Study, year 2007 to 2009), and Chinese (the Singapore Chinese Eye Study, year 2009 to 2011). Details of the study design and methodology of the SEED study have been described in detailed previously.
      • Foong A.W.
      • Saw S.M.
      • Loo J.L.
      • et al.
      Rationale and methodology for a population-based study of eye diseases in Malay people: The Singapore Malay eye study (SiMES).
      • Lavanya R.
      • Jeganathan V.S.
      • Zheng Y.
      • et al.
      Methodology of the Singapore Indian Chinese Cohort (SICC) eye study: quantifying ethnic variations in the epidemiology of eye diseases in Asians.
      In brief, the SEED study was conducted in the southwestern part of Singapore (Figure 1), using a standardized study protocol across the 3 ethnic groups of subjects. According to the Singapore census of 2000, the southwest region was unique in providing a fair distribution of the local population in terms of age, housing types, ethnic diversity, and socioeconomic status. In addition, the southwestern part of Singapore was selected mainly because of the following potential advantages. First, in this area, there were sufficient numbers of Malay, Indian, and Chinese individuals to obtain our original intended sample size. Second, the selected districts lay along the track line of Singapore's subway train, which allowed for a direct commute to the study clinic. Third, the study area covered approximately 110.4 sq km (42.6 sq mile) or 15.8% of the country's total land area of 699.4 sq km (270.0 sq mile), which covered a large sector of all populated areas in Singapore.
      Figure thumbnail gr1
      Figure 1Study sampling area of the Singapore Epidemiology of Eye Diseases Study.
      Age-stratified random sampling strategy was adopted in each ethnic group to select adults aged 40 to 80 years. A total of 5600 Malay residents, 6350 Indian residents, and 6752 Chinese residents were originally selected from Singapore's Ministry of Home Affairs. Ethnicity was defined and categorized based on the ethnicity indicated on each individual's National Registration Identity Card. From here, a list of eligible individuals was then identified. Ineligible individuals were defined as those who had permanently moved out from the stated residential address, had not lived at the given address in the past 6 months, or were deceased or terminally ill. A total of 4168 Malay, 4497 Indian, and 4605 Chinese individuals were identified as eligible individuals and invited to participate in the study; of which 3280 Malay (78.7% response rate), 3400 Indian (75.6% response rate), and 3353 Chinese (72.8%) individuals participated and underwent the study examinations. In each ethnic cohort, nonparticipants on average were slightly older than participants (P < .001 for all), but there were no significant sex differences between the 2 groups (P > .05 for all). Details on the sampling and enrollment of subjects are illustrated in Figure 2. The study was approved by the Singapore Eye Research Institute Institutional Review Board. All participants gave a written informed consent and the conduct of the study adhered to the Declaration of Helsinki. All subjects underwent standardized ocular, systemic examinations and interview at the Singapore Eye Research Institute.
      Figure thumbnail gr2
      Figure 2Selection and exclusion processes to form the final study population of the Singapore Epidemiology of Eye Diseases Study.

       Ocular Examination and Visual Acuity Testing

      All study participants underwent a detailed ocular examination, including visual acuity (VA) measurement, subjective refraction, slit lamp examination, intraocular pressure measurement, dilated fundus examination, and fundus photography. Digital fundus photography was performed using a 45-degree digital retinal camera (Canon CR DGi with a SLR digital camera back; Canon, Tokyo, Japan) after pupil dilation. Macular- and optic disc–centred fundus photos were captured for each eye.
      The presenting VA (PVA) with habitual correction, and best-corrected VA (BCVA) after subjective refraction were recorded using an Early Treatment of Diabetic Retinopathy Study logarithm of Minimum Angle of Resolution number chart (Lighthouse International) at a distance of 4 m.
      • Foong A.W.
      • Saw S.M.
      • Loo J.L.
      • et al.
      Rationale and methodology for a population-based study of eye diseases in Malay people: The Singapore Malay eye study (SiMES).
      When no numbers could be read at 4 m, the participant was moved to 3, 2, or 1 m, consecutively. When no number could be read at even 1 m, VA was then assessed as counting fingers, hand movements, perception of light, or no perception of light.

       Definitions of VI and Blindness

      Overall, both the United States and the modified WHO definitions were used to define VI and blindness. According to the US definition, VI was defined as VA <20/40 but ≥20/200 in the better-seeing eye, and blindness was defined as VA <20/200 in the better-seeing eye. Based on the WHO definition, VI was defined as VA <20/60 but ≥20/400 in the better-seeing eye, and blindness was defined as VA <20/400 in the better-seeing eye. We used a modified definition to classify individuals with VA of counting fingers or worse as blind. All subjects were also further categorized into 5 mutually exclusive categories: (1) bilateral blindness; (2) bilateral VI (VI in one eye and VI/blindness in the other eye); (3) unilateral blindness; (4) unilateral VI; and (5) normal vision in both eyes. Of note, unilateral VI and unilateral blindness cases were defined based on the worse-seeing eye, with the fellow eye having normal vision.

       Assessment on Causes of VI and Blindness

      Primary causes of VI or blindness were ascertained on the basis of clinical history, examination, disease definition, and clinical judgment. Undercorrected refractive error was defined when BCVA was at least 2 lines (in logarithm of Minimum Angle of Resolution chart) better than PVA. Glaucoma was diagnosed according to the International Society of Geographical and Epidemiological Ophthalmology scheme.
      • Foster P.J.
      • Buhrmann R.
      • Quigley H.A.
      • Johnson G.J.
      The definition and classification of glaucoma in prevalence surveys.
      AMD was graded from retinal photographs using the Wisconsin Age-related Maculopathy grading system.
      • Klein R.
      • Davis M.D.
      • Magli Y.L.
      • Segal P.
      • Klein B.E.
      • Hubbard L.
      The Wisconsin age-related maculopathy grading system.
      DR was graded from retinal photographs using a modification of the Arlie House classification system for the Early Treatment Diabetic Retinopathy Study.
      Grading diabetic retinopathy from stereoscopic color fundus photographs--an extension of the modified Airlie House classification. ETDRS report number 10. Early Treatment Diabetic Retinopathy Study Research Group.
      Cataract was determined from slit lamp examination and graded based on the using the Lens Opacities Classification System III.
      • Chylack Jr., L.T.
      • Wolfe J.K.
      • Singer D.M.
      • et al.
      The Lens Opacities Classification System III. The Longitudinal Study of Cataract Study Group.
      All retinal photographs were sent to the University of Sydney for grading of retinopathy, cataract, and other retinal diseases. These were all graded by a single grader, with adjudication performed by 2 experts in retinal diseases and photo grading, when necessary (Prof Paul Mitchell and Prof Jiejin Wang, collaborators from the University of Sydney).
      If there was more than 1 condition in the same eye, the main cause of VI and blindness was further determined, based on the severity of the underlying eye diseases. For example, in the case of intermediate cataract and mild DR, cataract would be indicated as the primary cause of VI.

       Questionnaire and Systemic Measurements

      Detailed interviewer-administered questionnaires were also performed to collect relevant medical history, socioeconomic status information (eg, education, income levels, type of housing), and lifestyle-related information (eg, cigarette use, alcohol consumption), as reported previously.
      • Lavanya R.
      • Jeganathan V.S.
      • Zheng Y.
      • et al.
      Methodology of the Singapore Indian Chinese Cohort (SICC) eye study: quantifying ethnic variations in the epidemiology of eye diseases in Asians.
      Cognitive assessment was performed for participants aged 60 and older, using the Abbreviated Mental Test (AMT), which consists of 10 questions of general cognitive function. Cognitive impairment was defined as an AMT score of 6 or less of 10 for the participants with 0 to 6 years of formal education, and an AMT score of 8 or less of 10 for those with more than 6 years of formal education.
      • Ong S.Y.
      • Cheung C.Y.
      • Li X.
      • et al.
      Visual impairment, age-related eye diseases, and cognitive function: the Singapore Malay Eye study.
      Deafness was defined based on self-reported history of hearing loss.
      Systolic and diastolic blood pressures (BPs) were measured using an automated sphygmomanometer (DinamapPro100V2; GE HealthCare, Little Chalfont, UK). BP was measured twice, with 5 minutes apart. A third measurement was taken if the previous 2 systolic BP readings differed by more than 10 mm Hg or the diastolic BP by more than 5 mm Hg. The mean between the 2 BP closest readings was then taken. Nonfasting blood samples were extracted from participants to determine levels of serum glucose, glycosylated haemoglobin (HbA1C), cholesterol, and serum creatinine. Body mass index (BMI) also was measured. Patients with hypertension were defined as having systolic BP ≥140 mm Hg, diastolic BP ≥90 mm Hg, use of antihypertensive medications, or self-reported physician-diagnosed hypertension. Diabetes was defined as having random glucose level of ≥11.1 mmol/L, HbA1c of ≥6.5 mmol/L, use of diabetic medications, or self-reported physician-diagnosed diabetes. Hyperlipidemia was defined as total cholesterol ≥6.2 mmol/L or use of lipid-lowering drugs. Cardiovascular disease (CVD) was defined as self-reported myocardial infarction or angina or stroke. Chronic kidney disease (CKD) was defined as having an estimated glomerular filtration rate <60 mL/min per 1.73 m2, based on the US National Kidney Foundation Kidney Disease Outcome Quality Initiative Working Group definition.
      K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification.
      Estimated glomerular filtration rate was estimated from serum creatinine level, using the CKD Epidemiology Collaboration equation.
      • Levey A.S.
      • Stevens L.A.
      • Schmid C.H.
      • et al.
      A new equation to estimate glomerular filtration rate.
      Systemic comorbidities were defined as concurrent presence of hypertension, diabetes, hyperlipidemia, CKD, or CVD. BMI was measured as weight in kilograms divided by height in meters squared. BMI was further divided into subgroups based on the WHO classification: underweight (BMI <18.5), normal (18.5 ≤ BMI <25), overweight (25 ≤ BMI <30), or obese (BMI ≥30).

       Statistical Analysis

      All statistical analyses were performed using R version 2.15.3 (R Development Core Team, 2013, Vienna, Austria). In descriptive analyses, to compare characteristics across the 3 ethnic groups, 1-way analysis of variance was performed for continuous variables, and χ2 tests were used for categorical variables. To compare characteristics between individuals with VI and those without, independent t-test was performed for continuous variables, and χ2 test was used for categorical variables. Prevalence estimates of VI and blindness for entire sample and respective ethnic groups were calculated and standardized to the 2010 Singapore Population Census 2010. Bootstrapping was performed to compare the standardized rate of VI and blindness across the 3 ethnicities, between sex groups, and between diabetic and nondiabetic groups.
      Multiple logistic regression model was used to assess the associations among demographic, socioeconomic, and systemic factors with presenting and best-corrected bilateral visual loss (defined as bilateral VI or blindness, based on better-seeing eye and US definition) in overall SEED sample and respective ethnic groups. Statistical interactions among demographic, socioeconomic, and systemic factors were examined in separate models by including cross-product interaction terms in the corresponding logistic regression models. Furthermore, for factors that were statistically significant (P < .05), the adjusted odds ratio (OR) as determined from the multiple logistic regression models and the prevalence of exposure factors were used to calculate the population of attributable risk (PAR) for these factors.
      • Bruzzi P.
      • Green S.B.
      • Byar D.P.
      • Brinton L.A.
      • Schairer C.
      Estimating the population attributable risk for multiple risk factors using case-control data.
      Last, factors that were significant (as determined from the multiple logistic regression model) in at least 2 ethnic groups were further illustrated in bar graphs for comparison of PAR across ethnic groups.

      Results

      Of the total 10 033 study participants, 13 did not have presenting VA measured, thus leaving 10 020 participants (3269 Malay, 3400 Indian, 3351 Chinese) included for presenting VA-related analysis. In addition, another 34 participants did not have subjective refraction taken, leaving 9986 participants (3235 Malay, 3400 Indian, 3351 Chinese) included for best-corrected VA-related analysis. Table 1 shows the characteristics of the overall study sample, and respective ethnic groups. The mean (SD) age of the overall sample was 58.9 (10.4) years. Among the 3 ethnic groups, Malay individuals were slightly older, had higher BMI profile, had higher prevalence of hypertension and CKD, and had higher proportions of current smokers and individuals of low socioeconomic status. On the other hand, Indian individuals had higher prevalence of diabetes and CVD.
      Table 1Demographic, Systemic, and Socioeconomic Characteristics of Participants in the Singapore Epidemiology of Eye Diseases Study.
      Overall (N = 10 020)Malay (n = 3269)Indian (n = 3400)Chinese (n = 3351)P
      P value was based on analysis of variance test for continuous variables and chi-square tests for categorical variables.
      Age (y)58.9 (10.4)59.2 (11.0)57.8 (10.1)59.7 (9.9)< .001
      Female gender, n (%)5082 (50.7)1698 (51.9)1694 (49.8)1690 (50.4).206
      BMI, kg/m225.4 (4.7)26.4 (5.1)26.2 (4.8)23.7 (3.7)< .001
      BMI categories, n (%)
       Underweight (BMI < 18.5)434 (4.4)140 (4.3)94 (2.8)200 (6.0)< .001
       Normal (18.5 ≤ BMI < 25)4697 (47.2)1243 (38.3)1381 (40.8)2073 (62.3)
       Overweight (25 ≤ BMI <30)3393 (34.1)1184 (36.5)1330 (39.3)879 (26.4)
       Obese (BMI ≥30)1435 (14.4)680 (20.9)578 (17.1)177 (5.3)
      Diabetes, n (%)2957 (29.5)1045 (32.0)1320 (38.8)592 (17.7)< .001
      Hypertension, n (%)6383 (63.9)2289 (70.3)2050 (60.4)2044 (61.1)< .001
      History of CVD, n (%)1081 (10.8)364 (11.2)483 (14.3)234 (7.0)< .001
      CKD, n (%)1213 (12.7)715 (22.8)273 (8.4)225 (7.1)< .001
      Hyperlipidaemia, n (%)4403 (45.4)1323 (41.2)1554 (47.9)1526 (47.1)< .001
      Education level, n (%)
       Formal education7676 (76.7)2295 (70.3)2786 (82.1)2595 (77.5)< .001
       No formal education2331 (23.3)968 (29.7)608 (17.9)755 (22.5)
      Month income status, n (%)
       Income ≥ S$20002187 (22.3)348 (10.7)870 (26.3)969 (29.8)< .001
       Income < S$20007616 (77.7)2893 (89.3)2441 (73.7)2282 (70.2)
      Current smoker, n (%)
       Never smoked or past smoker8402 (84.0)2598 (79.7)2895 (85.3)2909 (86.8)< .001
       Current smoker1603 (16.0)662 (20.3)499 (14.7)442 (13.2)
      Alcohol consumption, n (%)848 (8.5)53 (1.6)429 (12.6)366 (10.9)< .001
      Living alone, n (%)495 (5.0)157 (4.8)169 (5.0)169 (5.1)< .001
      Housing category (%)
       ≥ 5 room public housing flat3064 (30.6)517 (15.8)1211 (35.7)1336 (39.9)< .001
       3–4 room public housing flat6204 (62.0)2245 (68.8)2021 (59.6)1938 (57.9)
       1–2 room public housing flat738 (7.4)503 (15.4)161 (4.8)74 (2.2)
      Low socioeconomic status, n (%)
      Defined as having primary or lower education, individual monthly income < SGD$2000, and residing in 1–2 room public housing flat.
      627 (6.4)437 (13.5)132 (4.0)58 (1.8)< .001
      BMI = body mass index; CVD = cardiovascular disease; CKD = chronic kidney disease.
      Data presented are mean (standard deviation) or frequency (percentage), where appropriate.
      a P value was based on analysis of variance test for continuous variables and chi-square tests for categorical variables.
      b Defined as having primary or lower education, individual monthly income < SGD$2000, and residing in 1–2 room public housing flat.
      For the whole SEED sample, based on PVA and US definition, the age-standardized prevalence of bilateral blindness was 0.9% (95% confidence interval [CI], 0.7%–1.1%), 18.4% (95% CI, 17.6%–19.2%) for bilateral VI, 1.8% (95% CI, 1.6%–2.1%) for unilateral blindness, and 18.5% (95% CI, 17.7%–19.4%) for unilateral VI (Table 2). Based on BCVA and US definition, the age-standardized prevalence of bilateral blindness was 0.3% (95% CI, 0.2%–0.4%), 4.2% (95% CI, 3.8%–4.5%) for bilateral VI, 1.6% (95% CI, 1.4%–1.9%) for unilateral blindness, and 6.8% (95% CI, 6.3%–7.3%) for unilateral VI (Table 2). Overall, women had higher prevalence of bilateral VI and blindness (presenting and best corrected) compared with men. Similar trends were observed for VI and blindness based on WHO definition (albeit different rates than the ones based on US definition, Supplementary Table 2). For all categories of VI and blindness, the prevalence rates increased with older age groups (all P trend < .001, not shown in tables). This trend is particularly prominent for presenting and best-corrected bilateral VI, with exponential increase in prevalence rates from age group 60 to 69 to age group ≥70 years.
      Table 2Prevalence of Blindness and VI in the Singapore Epidemiology of Eye Diseases Study, Based on the US Definition
      Vision status
      Bilateral blindness and bilateral VI were defined based on better-seeing eye, whereas unilateral blindness and unilateral VI were defined based on worse-seeing eye.
      Age group, yBased on Presenting Visual AcuityBased on Best-Corrected Visual Acuity
      All (N = 10 020)Male (n = 4938)Female (n = 5082)All (n = 9986)Male (n = 4927)Female (n = 5059)
      n (%)n (%)n (%)n (%)n (%)n (%)
      Bilateral blindness40–499 (0.4)1 (0.1)8 (0.6)2 (0.08)0 (0.0)2 (0.2)
      50–5914 (0.5)6 (0.4)8 (0.5)4 (0.1)3 (0.2)1 (0.06)
      60–6928 (1.1)11 (0.8)17 (1.4)11 (0.4)6 (0.5)5 (0.4)
      70+58 (3.2)16 (1.6)42 (4.9)23 (1.3)8 (0.8)15 (1.8)
      Total109 (1.1)34 (0.7)75 (1.5)40 (0.4)17 (0.4)23 (0.5)
      Age standardized prevalence, % (95% CI)
      Standardized to Singapore population 2010 census.
      0.9 (0.7–1.1)0.5 (0.4–0.8)1.3 (1.0–1.6)0.3 (0.2–0.4)0.3 (0.2–0.4)0.4 (0.2–0.6)
      Bilateral VI
      Bilateral VI includes participants with VI in one eye and VI/blindness in the other eye.
      40–49220 (8.9)80 (6.7)140 (10.9)15 (0.6)6 (0.5)9 (0.7)
      50–59455 (14.5)173 (11.9)282 (16.8)43 (1.4)9 (0.6)34 (2.0)
      60–69655 (25.6)284 (21.7)371 (29.6)129 (5.1)47 (3.6)82 (6.6)
      70+823 (44.9)409 (42.0)414 (48.1)357 (19.6)148 (15.3)209 (24.6)
      Total2153 (21.5)946 (19.2)1207 (23.8)544 (5.5)210 (4.3)334 (6.6)
      Age standardized prevalence, % (95% CI)
      Standardized to Singapore population 2010 census.
      18.4 (17.6–19.2)15.7 (14.6–16.8)21.0 (19.8–22.3)4.2 (3.8–4.5)3.0 (2.6–3.5)5.3 (4.8–5.9)
      Unilateral blindness40–4927 (1.1)17 (1.4)10 (0.8)14 (0.6)9 (0.8)5 (0.4)
      50–5941 (1.3)18 (1.2)23 (1.4)39 (1.2)17 (1.2)22 (1.3)
      60–6965 (2.5)38 (2.9)27 (2.2)55 (2.2)32 (2.4)23 (1.9)
      70+73 (4.0)43 (4.4)30 (3.5)86 (4.7)53 (5.5)33 (3.9)
      Total206 (2.1)116 (2.4)90 (1.8)194 (1.9)111 (2.3)83 (1.6)
      Age standardized prevalence, % (95% CI)
      Standardized to Singapore population 2010 census.
      1.8 (1.6–2.1)2.0 (1.7–2.5)1.6 (1.3–2.0)1.6 (1.4–1.9)1.8 (1.5–2.2)1.4 (1.1–1.8)
      Unilateral VI40–49319 (12.9)160 (13.4)159 (12.4)47 (1.9)25 (2.1)22 (1.7)
      50–59588 (18.7)268 (18.4)320 (19.0)118 (3.8)49 (3.4)69 (4.1)
      60–69661 (25.8)367 (28.0)294 (23.5)295 (11.6)151 (11.5)144 (11.6)
      70+444 (24.2)246 (25.2)198 (23.0)389 (21.4)197 (20.4)192 (22.6)
      Total2012 (20.1)1041 (21.1)971 (19.1)849 (8.5)422 (8.6)427 (8.4)
      Age standardized prevalence, % (95% CI)
      Standardized to Singapore population 2010 census.
      18.5 (17.7–19.4)19.2 (17.9–20.4)17.9 (16.7–19.1)6.8 (6.3–7.3)6.6 (5.9–7.3)7.0 (6.3–7.7)
      CI = confidence interval; VI = visual impairment.
      Based on the US definition, VI was defined as VA <20/40 to ≥20/200. Blindness was defined as VA <20/200.
      a Bilateral blindness and bilateral VI were defined based on better-seeing eye, whereas unilateral blindness and unilateral VI were defined based on worse-seeing eye.
      b Standardized to Singapore population 2010 census.
      c Bilateral VI includes participants with VI in one eye and VI/blindness in the other eye.
      Table 3 shows the prevalence of blindness and VI, based on US definition, across the 3 ethnic groups. For presenting bilateral blindness, we observed that Malay individuals (1.4%) had higher age-standardized prevalence, compared with Indian (0.7%, P < .001) and Chinese (0.5%, P < .001) individuals. Malay individuals also had higher age-standardized prevalence of presenting bilateral VI (19.9%), compared with Indian (18.0%) and Chinese (17.2%) individuals. These ethnic differences in prevalence rates were consistently observed for all age groups. However, Indian and Chinese individuals (both 2.1%) had higher prevalence of unilateral blindness compared with Malay individuals (1.3%, both P < .001). Chinese individuals had higher prevalence of unilateral VI (20.9%) compared with Indian (18.0%, P < .001) and Malay individuals (16.5%, P < .001). On the other hand, based on BCVA, the prevalence of best-corrected bilateral blindness was similar across the 3 ethnicities. For best-corrected bilateral VI, Malay individuals (5.4%) had higher age-standardized prevalence, compared with Indian (3.6%, P < .001) and Chinese individuals (3.3%, P < .001). Indian individuals had higher prevalence of best-corrected unilateral blindness, compared with Malay individuals (1.1%, P < .001). Chinese individuals had the higher prevalence of unilateral VI (7.8%) compared with Indian (6.2%, P < .001) and Malay individuals (6.2%, P < .001). Supplementary Table 3 shows the prevalence of blindness and VI across the 3 ethnic groups, based on WHO definition. In general, the trends in ethnic differences were similar to the trends based on US definition (Table 3).
      Table 3Prevalence of Blindness and Visual Impairment (VI) in the Three Ethnic Groups of the Singapore Epidemiology of Eye Diseases Study, Based on the United States Definition
      Vision status
      Bilateral blindness and bilateral VI were defined based on better-seeing eye, whereas unilateral blindness and unilateral VI were defined based on worse-seeing eye.
      Age group, yBased on Presenting Visual AcuityBased on Best-Corrected Visual Acuity
      Malay (n = 3269)Indian (n = 3400)Chinese (n = 3351)Malay (n = 3235)Indian (n = 3400)Chinese (n = 3351)
      n (%)n (%)n (%)n (%)n (%)n (%)
      Bilateral blindness40–495 (0.6)3 (0.3)1 (0.1)0 (0.0)2 (0.2)0 (0.0)
      50–594 (0.4)5 (0.5)5 (0.5)0 (0.0)3 (0.3)1 (0.1)
      60–6914 (1.8)10 (1.1)4 (0.4)5 (0.7)4 (0.5)2 (0.2)
      70+37 (5.1)9 (1.9)12 (1.9)12 (1.7)6 (1.3)5 (0.8)
      Total60 (1.8)27 (0.8)22 (0.7)17 (0.5)15 (0.4)8 (0.2)
      Age standardized prevalence, % (95% CI)
      Standardized to Singapore population 2010 census.
      1.4 (1.0–1.8)0.7 (0.5–1.1)0.5 (0.3–0.9)0.3 (0.2–0.6)0.4 (0.2–0.7)0.2 (0.1–0.4)
      Bilateral VI
      Bilateral VI includes participants with VI in one eye and VI/blindness in the other eye.
      40–4977 (9.5)81 (8.5)62 (8.7)7 (0.9)5 (0.5)3 (0.4)
      50–59156 (16.3)174 (16.2)125 (11.3)19 (2.0)18 (1.7)6 (0.5)
      60–69212 (27.3)233 (26.3)210 (23.3)49 (6.4)51 (5.8)29 (3.2)
      70+344 (47.5)181 (37.8)298 (47.2)179 (25.3)66 (13.8)112 (17.8)
      Total789 (24.1)669 (19.7)695 (20.7)254 (7.9)140 (4.1)150 (4.5)
      Age standardized prevalence, % (95% CI)
      Standardized to Singapore population 2010 census.
      19.9 (18.4–21.4)18.0 (16.6–19.4)17.2 (15.9–18.7)5.4 (4.8–6.2)3.6 (3.0–4.2)3.3 (2.7–3.9)
      Unilateral blindness40–492 (0.3)13 (1.4)12 (1.7)0 (0.0)9 (0.9)5 (0.7)
      50–598 (0.8)16 (1.5)17 (1.5)7 (0.7)21 (2.0)11 (1.0)
      60–6920 (2.6)20 (2.3)25 (2.8)16 (2.1)21 (2.4)18 (2.0)
      70+25 (3.5)27 (5.6)21 (3.3)26 (3.7)30 (6.3)30 (4.8)
      Total55 (1.7)76 (2.2)75 (2.2)49 (1.5)81 (2.4)64 (1.9)
      Age standardised prevalence, % (95% CI)
      Standardized to Singapore population 2010 census.
      1.3 (1.0–1.7)2.1 (1.7–2.7)2.1 (1.6–2.6)1.1 (0.8–1.5)2.2 (1.8–2.8)1.6 (1.2–2.1)
      Unilateral VI40–4986 (10.6)120 (12.5)113 (15.9)7 (0.9)24 (2.5)16 (2.3)
      50–59156 (16.3)195 (18.1)237 (21.4)35 (3.7)44 (4.1)39 (3.5)
      60–69202 (26.0)212 (23.9)247 (27.4)93 (12.1)89 (10.0)113 (12.6)
      70+168 (23.2)119 (24.8)157 (24.9)137 (19.4)78 (16.3)174 (27.6)
      Total612 (18.7)646 (19.0)754 (22.5)272 (8.4)235 (6.9)342 (10.2)
      Age standardized prevalence, % (95% CI)
      Standardized to Singapore population 2010 census.
      16.8 (15.5–18.3)18.0 (16.6–19.5)20. 9 (19.3–22.6)6.2 (5.5–7.1)6.2 (5.4–7.1)7.8 (7.0–8.8)
      CI = confidence interval; VI = visual impairment.
      Based on the US definition, VI was defined as VA <20/40 to ≥20/200. Blindness was defined as VA <20/200.
      a Bilateral blindness and bilateral VI were defined based on better-seeing eye, whereas unilateral blindness and unilateral VI were defined based on worse-seeing eye.
      b Standardized to Singapore population 2010 census.
      c Bilateral VI includes participants with VI in one eye and VI/blindness in the other eye.
      Tables 4 and 5 show the causes of VI and blindness (US definition) in the SEED study for overall sample and respective ethnic groups. Based on PVA (Table 4), the leading causes of bilateral VI were undercorrected refractive error (1202 cases; 55.8%), cataract (735 cases; 34.1%), DR (56 cases; 2.6%), and AMD (33 cases; 1.5%). This distribution was similarly observed across the 3 ethnic groups (Figure 3). On the other hand, for presenting bilateral blindness, the leading causes were cataract (67 cases; 61.5%), undercorrected refractive error (17 cases; 15.6%), AMD (9 cases; 8.3%), and DR (5 cases; 4.6%). This was also similarly observed across the 3 ethnic groups, except for Indian individuals, in whom AMD was the second leading cause instead (Figure 4). After refractive correction (ie, based on BCVA, Table 5), the leading causes of bilateral VI were cataract (422 cases; 77.6%), DR (35 cases; 6.4%), myopic maculopathy (18 cases; 3.3%), and posterior capsular opacity (17 cases; 3.1%). Across the 3 ethnic groups, cataract and DR were consistently the top 2 causes of best-corrected bilateral VI (Figure 5). Of note, there was a disproportionately higher number of cataract cases in Malay individuals (n = 217; 85.4%), compared with Indian (n = 91; 65%) and Chinese individuals (n = 114; 76%). In addition, myopic maculopathy was the third leading cause for Indian (n = 9) and Chinese (n = 7 individuals), whereas there were only 2 cases of myopic maculopathy in Malay individuals. For best-corrected bilateral blindness, cataract remained to be the primary cause (65%); this was also consistent across the 3 ethnic groups (Figure 6).
      Table 4Causes of VI and Blindness in the Singapore Epidemiology of Eye Diseases Study, Based on Presenting Visual Acuity and the United States Definition.
      CausesBilateral VI
      Bilateral VI includes participants with VI in one eye and VI/blindness in the fellow eye.
      Bilateral Blindness
      Overall SEED (n = 2153)Malay (n = 789)Indian (n = 669)Chinese (n = 695)Overall SEED (n = 109)Malay (n = 60)Indian (n = 27)Chinese (n = 22)
      n (%)n (%)n (%)n (%)n (%)n (%)n (%)n (%)
      Undercorrected refractive error1202 (55.8)458 (58.0)337 (50.4)407 (58.6)17 (15.6)10 (16.7)2 (7.4)5 (22.7)
      Cataract735 (34.1)278 (35.2)226 (33.8)231 (33.2)67 (61.5)42 (70.0)16 (59.3)9 (40.9)
      Diabetic retinopathy56 (2.6)17 (2.2)29 (4.3)10 (1.4)5 (4.6)2 (3.3)1 (3.7)2 (9.1)
      Age-related macular degeneration33 (1.5)8 (1.0)15 (2.2)10 (1.4)9 (8.3)2 (3.3)3 (11.1)4 (18.2)
      Posterior capsular opacity28 (1.3)4 (0.5)15 (2.2)9 (1.3)0 (0)0 (0)0 (0)0 (0)
      Myopic maculopathy19 (0.9)3 (0.4)11 (1.6)5 (0.7)1 (0.9)0 (0)0 (0)1 (4.5)
      Glaucoma16 (0.7)6 (0.8)5 (0.7)5 (0.7)3 (2.8)2 (3.3)1 (3.7)0 (0)
      Epiretinal membrane7 (0.3)0 (0)2 (0.3)5 (0.7)0 (0)0 (0)0 (0)0 (0)
      Other maculopathy13 (0.6)1 (0.1)12 (1.8)0 (0)2 (1.8)0 (0)2 (7.4)0 (0)
      Corneal opacity/scar13 (0.6)2 (0.3)9 (1.3)2 (0.3)2 (1.8)1 (1.7)0 (0)1 (4.5)
      Pterygium3 (0.1)2 (0.3)0 (0)1 (0.1)0 (0)0 (0)0 (0)0 (0)
      Retinal detachment3 (0.1)1 (0.1)1 (0.1)1 (0.1)2 (1.8)1 (1.7)1 (0.9)0 (0)
      Amblyopia8 (0.4)2 (0.3)5 (0.7)1 (0.1)1 (0.9)0 (0)1 (3.7)0 (0)
      Retinal vein occlusion1 (0)1 (0.1)0 (0)0 (0)0 (0)0 (0)0 (0)0 (0)
      Ocular trauma2 (0.1)2 (0.3)0 (0)0 (0)0 (0)0 (0)0 (0)0 (0)
      Aphakia1 (0)1 (0.1)0 (0)0 (0)0 (0)0 (0)0 (0)0 (0)
      Others13 (0.6)3 (0.3)2 (0.3)8 (1.2)0 (0)0 (0)0 (0)0 (0)
      SEED = Singapore Epidemiology of Eye Disease; VI = visual impairment.
      a Bilateral VI includes participants with VI in one eye and VI/blindness in the fellow eye.
      Table 5Causes of VI and Blindness in the Singapore Epidemiology of Eye Diseases Study, Based on Best-corrected Visual Acuity and the US Definition.
      CausesBilateral VI
      Bilateral VI includes participants with VI in one eye and VI/blindness in the fellow eye.
      Bilateral Blindness
      Overall SEED (n = 544)Malay (n = 254)Indian (n = 140)Chinese (n = 150)Overall SEED (n = 40)Malay (n = 17)Indian (n = 15)Chinese (n = 8)
      n (%)n (%)n (%)n (%)n (%)n (%)n (%)n (%)
      Cataract422 (77.6)217 (85.4)91 (65.0)114 (76.0)26 (65.0)13 (76.5)9 (60.0)4 (50.0)
      Diabetic retinopathy35 (6.4)11 (4.3)16 (11.4)8 (5.3)2 (5.0)0 (0)1 (6.7)1 (12.5)
      Age-related macular degeneration13 (2.4)7 (2.8)2 (1.4)4 (2.7)6 (15.0)1 (5.9)3 (20.0)2 (25.0)
      Posterior capsular opacity17 (3.1)3 (1.2)7 (5.0)7 (4.7)0 (0)0 (0)0 (0)0 (0)
      Myopic maculopathy18 (3.3)2 (0.8)9 (6.4)7 (4.7)0 (0)0 (0)0 (0)0 (0)
      Glaucoma8 (1.5)4 (1.6)4 (2.9)0 (0)2 (5.0)2 (11.8)0 (0)0 (0)
      Epiretinal membrane5 (0.9)0 (0)1 (0.7)4 (2.7)0 (0)0 (0)0 (0)0 (0)
      Other maculopathy2 (0.4)0 (0)2 (1.4)0 (0)1 (2.5)0 (0)1 (6.7)0 (0)
      Corneal opacity/scar7 (1.3)2 (0.8)5 (3.6)0 (0)1 (2.5)0 (0)0 (0)1 (12.5)
      Pterygium2 (0.4)1 (0.4)0 (0)1 (0.7)0 (0)0 (0)0 (0)0 (0)
      Retinal detachment3 (0.6)1 (0.4)1 (0.7)1 (0.7)2 (5.0)1 (5.9)1 (6.7)0 (0)
      Amblyopia6 (1.1)2 (0.8)2 (1.4)2 (1.3)0 (0)0 (0)0 (0)0 (0)
      Retinal vein occlusion1 (0.2)0 (0)0 (0)1 (0.7)0 (0)0 (0)0 (0)0 (0)
      Ocular trauma2 (0.4)2 (0.8)0 (0)0 (0)0 (0)0 (0)0 (0)0 (0)
      Aphakia0 (0)0 (0)0 (0)0 (0)0 (0)0 (0)0 (0)0 (0)
      Others3 (0.6)2 (0.8)0 (0)1 (0.7)0 (0)0 (0)0 (0)0 (0)
      SEED = Singapore Epidemiology of Eye Disease; VI = visual impairment.
      a Bilateral VI includes participants with VI in one eye and VI/blindness in the fellow eye.
      Figure thumbnail gr3
      Figure 3Comparison of top causes of presenting bilateral visual impairment (VI, US definition), across the 3 ethnicities in the Singapore Epidemiology of Eye Diseases Study. AMD= age-related macular degeneration; DR= diabetic retinopathy.
      Figure thumbnail gr4
      Figure 4Comparison of top causes of presenting bilateral blindness (US definition), across the 3 ethnicities in the Singapore Epidemiology of Eye Diseases Study. AMD = age-related macular degeneration; DR = diabetic retinopathy.
      Figure thumbnail gr5
      Figure 5Comparison of top causes of best-corrected bilateral visual impairment (VI, US definition), across the 3 ethnicities in the Singapore Epidemiology of Eye Diseases Study. AMD = age-related macular degeneration; DR = diabetic retinopathy.
      Figure thumbnail gr6
      Figure 6Comparison of top causes of best-corrected bilateral blindness (US definition), across the 3 ethnicities in the Singapore Epidemiology of Eye Diseases Study. AMD= age-related macular degeneration; DR= diabetic retinopathy.
      Table 6 shows demographic, systemic, lifestyle, and socioeconomic factors associated with presenting bilateral visual loss (based on US definition) in the SEED study. Multiple logistic regression model demonstrated that older age (per decade, OR, 1.85; 95% CI, 1.73–1.98), female sex (OR, 1.42; 95% CI, 1.25–1.62), presence of CKD (OR, 1.22; 95% CI, 1.05–1.43), cognitive impairment (OR, 2.15; 95% CI, 1.75–2.63), smaller housing type (those residing in 1- to 2-room public flats; OR, 1.88; 95% CI, 1.50–2.36), lower income (monthly income less than SGD $2,000; OR, 1.47; 95% CI, 1.23–1.77), and no formal education (OR, 1.75; 95% CI, 1.54–1.99) were significantly associated with higher risk of presenting bilateral VI or blindness. Of these factors, older age, female sex, cognitive impairment, smaller housing type, and lower education level were consistently significant across the 3 ethnic groups (all P ≤ .046). However, CKD was significant only in Indian individuals (OR, 1.58; P = .003) but not Malay and Chinese individuals. Of note, in this multiple regression model adjusting for multiple factors, ethnicity was not significantly associated with risk of presenting bilateral visual loss (all P ≥ .225)
      Table 6Factors Associated With Presenting Bilateral Visual Loss (US Definition) in the Singapore Epidemiology of Eye Diseases Study
      Presenting Bilateral Visual Loss as Outcome
      Inclusive of bilateral visual impairment and bilateral blindness, based on the better-seeing eye.
      Overall Sample (N = 2262)
      In this model, ethnicity was also adjusted, in addition to covariates listed in table.
      Malay (n = 849)Indian (n = 696)Chinese (n = 717)
      OR (95% CI)POR (95% CI)POR (95% CI)POR (95% CI)P
      Age (per decade)1.85 (1.73–1.98)< .0012.01 (1.79–2.26)< .0011.61 (1.43–1.80)< .0012.04 (1.80–2.31)< .001
      Female1.42 (1.25–1.62)< .0011.82 (1.44–2.28)< .0011.27 (1.00–1.59).0461.28 (1.03–1.60).026
      Ethnicity
       ChineseReference
       Malay1.06 (0.92–1.23).412
       Indian1.09 (0.95–1.26).225
      Medical conditions
       Hypertension0.99 (0.87–1.13).8861.13 (0.88–1.45).3240.92 (0.73–1.15).4630.93 (0.74–1.18).576
       Diabetes mellitus1.02 (0.90–1.15).7910.96 (0.79–1.17).7201.09 (0.89–1.34).4041.00 (0.78–1.28).977
       Cardiovascular disease1.09 (0.92–1.29).3361.15 (0.87–1.52).3171.05 (0.80–1.38).7261.14 (0.80–1.63).474
       Chronic kidney disease1.22 (1.05–1.43).0111.01 (0.81–1.26).9261.58 (1.16–2.14).0031.26 (0.90–1.76).174
       Hyperlipidaemia0.98 (0.87–1.10).7041.06 (0.87–1.27).5710.97 (0.79–1.18).7310.88 (0.72–1.08).218
       Systemic comorbidities
      Systemic comorbidities were classified based on the concurrent presence of 2 or more systemic conditions, namely, diabetes, hyperlipidemia, hypertension, chronic kidney disease, or cardiovascular disease. Evaluation of systemic comorbidities as exposure variable was performed in a separate model, where individual exposures of diabetes, hypertension, hyperlipidemia, chronic kidney disease, and cardiovascular disease were not included in the model.
      No systemic diseaseReferenceReferenceReferenceReference
      Any 1 systemic disease0.93 (0.78–1.11).4401.03 (0.73–1.44).8771.07 (0.79–1.46).6440.75 (0.56–1.00).052
      Any 2 systemic diseases0.98 (0.82–1.18).8211.12 (0.79–1.60).5201.15 (0.84–1.57).3740.71 (0.53–0.97).029
      ≥ 3 systemic diseases1.03 (0.85–1.24).7751.12 (0.78–1.59).5481.16 (0.85–1.59).3550.82 (0.59–1.16).261
      P-trend.463.449.329.318
      Cognitive impairment
      Cognitive impairment was assessed in a separate model, including only subjects aged 60 years and older, while adjusting for the same set of covariates as original main model.
      2.15 (1.75–2.63)<.0012.80 (2.04–3.86)<.0011.93 (1.35–2.78)<.0011.52 (1.01–2.29).044
      Deaf1.86 (0.95–3.62).0702.72 (0.89–8.33).0802.41 (0.89–6.53).0830.61 (0.10–3.54).580
      BMI category
       Normal BMIReferenceReferenceReferenceReference
       Underweight1.20 (0.93–1.55).1631.16 (0.74–1.82).5261.22 (0.72–2.06).4541.17 (0.78–1.73).449
       Overweight0.87 (0.77–0.98).0270.83 (0.67–1.02).0810.85 (0.69–1.05).1410.90 (0.71–1.13).352
       Obese0.83 (0.70–0.99).0360.75 (0.58–0.97).0270.86 (0.65–1.15).3080.83 (0.53–1.30).414
      Current smokers1.15 (0.97–1.37).0961.13 (0.86–1.50).3811.16 (0.85–1.58).3431.26 (0.93–1.71).142
      Alcohol consumption1.06 (0.85–1.32).6150.61 (0.21–1.79).3691.12 (0.82–1.54).4700.94 (0.66–1.32).709
      Housing type
       ≥5 room public housing flatReferenceReferenceReferenceReference
       3–4 room public housing flat1.33 (1.16–1.52)< .0011.19 (0.89–1.59).2421.22 (0.98–1.52).0751.52 (1.22–1.90)< .001
       1–2 room public housing flat1.88 (1.50–2.36)< .0011.64 (1.15–2.33).0061.58 (1.01–2.46).0442.90 (1.64–5.14)< .001
      Living alone0.83 (0.65–1.05).1251.14 (0.75–1.73).5330.70 (0.45–1.10).1190.70 (0.45–1.08).107
      Monthly income category
       ≥ S$2000ReferenceReferenceReferenceReference
       < S$20001.47 (1.23–1.77)< .0011.56 (0.97–2.50).0671.36 (1.03–1.80).0291.62 (1.22–2.17).001
      Education category
       Formal educationReferenceReferenceReferenceReference
       No formal education1.75 (1.54–1.99)< .0011.64 (1.33–2.03)< .0011.80 (1.42–2.29)< .0011.66 (1.32–2.08)< .001
      BMI = body mass index; CI = confidence interval; OR= odds ratio.
      a Inclusive of bilateral visual impairment and bilateral blindness, based on the better-seeing eye.
      b In this model, ethnicity was also adjusted, in addition to covariates listed in table.
      c Systemic comorbidities were classified based on the concurrent presence of 2 or more systemic conditions, namely, diabetes, hyperlipidemia, hypertension, chronic kidney disease, or cardiovascular disease. Evaluation of systemic comorbidities as exposure variable was performed in a separate model, where individual exposures of diabetes, hypertension, hyperlipidemia, chronic kidney disease, and cardiovascular disease were not included in the model.
      d Cognitive impairment was assessed in a separate model, including only subjects aged 60 years and older, while adjusting for the same set of covariates as original main model.
      For best-corrected bilateral visual loss (Table 7), older age (per decade, OR, 2.90; 95% CI, 2.53–3.33), female sex (OR, 1.74; 95% CI, 1.37–2.20), Malay individuals (compared with Chinese; OR, 1.72; 95% CI, 1.32–2.23), presence of diabetes (OR, 1.43; 95% CI, 1.16–1.76), presence of CKD (OR, 1.38; 95% CI, 1.10–1.74), cognitive impairment (OR, 2.07; 95% CI, 1.60–2.68), ≥3 systemic comorbidities (OR, 2.05; 95% CI, 1.30–3.23), smaller housing type (OR, 1.51; 95% CI, 1.05–2.19), lower income (OR, 2.84; 95% CI, 1.47–5.49), and no formal education (OR, 1.81; 95% CI, 1.45–2.26) were significant associated risk factors. Of these factors, older age, female sex, cognitive impairment, lower education level, and diabetes (borderline significance for Malay individuals) were common risk factors across the 3 ethnic groups. On the other hand, CKD was significant only in Indian individuals (OR, 1.75; P = .02), and deafness was significant only in Malay individuals (OR, 5.51; P = .016). Systemic comorbidities were significant in both Malay and Indian individuals, with more prominent effect observed in Indian individuals.
      Table 7Factors Associated with Best-Corrected Bilateral Visual Loss (US Definition) in the Singapore Epidemiology of Eye Diseases Study
      Best-Corrected Bilateral Visual Loss as Outcome
      Inclusive of bilateral visual impairment and bilateral blindness, based on the better-seeing eye.
      Overall Sample (N = 584)
      In this model, ethnicity was also adjusted, in addition to covariates listed in the table.
      Malay (n = 849)Indian (n = 696)Chinese (n = 717)
      OR (95% CI)POR (95% CI)POR (95% CI)POR (95% CI)P
      Age (per decade)2.90 (2.53–3.33)< .0013.59 (2.86–4.51)< .0011.95 (1.54–2.46)< .0013.75 (2.83–4.96)< .001
      Female1.74 (1.37–2.20)< .0012.03 (1.41–2.91)< .0011.94 (1.20–3.12).0071.25 (0.81–1.93).308
      Ethnicity
       ChineseReference
       Malay1.72 (1.32–2.23)< .001
       Indian1.23 (0.93–1.63).149
      Medical conditions
       Hypertension1.14 (0.85–1.53).3891.06 (0.67–1.69).7951.57 (0.91–2.69).1051.00 (0.57–1.77).994
       Diabetes mellitus1.43 (1.16–1.76).0011.34 (0.99–1.81).0551.51 (1.01–2.25).0441.58 (1.02–2.44).039
       Cardiovascular disease1.08 (0.83–1.42).5601.10 (0.74–1.65).6311.03 (0.63–1.68).9041.23 (0.70–2.19).472
       Chronic kidney disease1.38 (1.10–1.74).0051.21 (0.89–1.65).2181.75 (1.09–2.82).0201.16 (0.69–1.92).577
       Hyperlipidaemia1.22 (0.99–1.50).0591.43 (1.06–1.94).0181.16 (0.77–1.75).4780.95 (0.63–1.43).793
       Systemic comorbidities
      Systemic comorbidities were classified based on the concurrent presence of 2 or more systemic conditions, namely, diabetes, hyperlipidemia, hypertension, chronic kidney disease, or cardiovascular disease. Evaluation of systemic comorbidities as exposure variable was performed in a separate model, where individual exposures of diabetes, hypertension, hyperlipidemia, chronic kidney disease, and cardiovascular disease were not included in the model.
      No systemic diseaseReferenceReferenceReferenceReference
      Any 1 systemic disease1.32 (0.83–2.10).2431.56 (0.66–3.67).312.25 (0.90–5.67).0840.82 (0.40–1.70).592
      Any 2 systemic diseases1.18 (0.74–1.89).4791.58 (0.68–3.70).292.23 (0.88–5.63).090.54 (0.26–1.15).113
      ≥ 3 systemic diseases2.05 (1.30–3.23).0022.37 (1.03–5.47).0434.03 (1.64–9.86).0021.12 (0.54–2.35).757
      P-trend<.001.004<.001.44
      Cognitive impairment
      Cognitive impairment was assessed in a separate model, including only subjects aged 60 years and older, while adjusting for the same set of covariates as original main model.
      2.07 (1.60–2.68)<.0011.85 (1.27–2.69).0012.52 (1.51–4.19)<.0012.12 (1.25–3.60).006
      Deaf2.06 (0.87–4.89).1015.51 (1.38–22.01).0161.53 (0.38–6.26).5521.27 (0.13–12.34).838
      BMI category
       Normal BMIReferenceReferenceReferenceReference
       Underweight1.31 (0.87–1.98).1921.29 (0.68–2.45).4331.63 (0.70–3.81).2591.06 (0.52–2.19).865
       Overweight0.68 (0.54–0.86).0010.76 (0.54–1.07).1120.58 (0.37–0.90).0140.66 (0.40–1.07).090
       Obese0.71 (0.52–0.97).0290.78 (0.51–1.17).2300.50 (0.28–0.90).0200.93 (0.39–2.22).868
      Current smokers1.22 (0.87–1.70).2441.13 (0.68–1.85).6431.37 (0.69–2.74).3681.37 (0.75–2.50).303
      Alcohol consumption1.00 (0.59–1.68).9910.68 (0.29–1.59).3771.37 (0.68–2.77).384
      Housing type
       ≥ 5 room public housing flatReferenceReferenceReferenceReference
       3–4 room public housing flat1.14 (0.87–1.49).3350.96 (0.58–1.57).8611.15 (0.73–1.81).5461.25 (0.78–1.98).351
       1–2 room public housing flat1.51 (1.05–2.19).0271.37 (0.78–2.39).2731.07 (0.48–2.37).8712.08 (0.84–5.16).116
      Living alone0.86 (0.57–1.30).4661.05 (0.58–1.88).8830.49 (0.19–1.32).1590.96 (0.43–2.10).910
      Monthly Income category
       ≥ S$2000ReferenceReferenceReferenceReference
       < S$20002.84 (1.47–5.49).0024.02 (0.54–29.95).1742.86 (1.11–7.41).0302.69 (0.93–7.76).067
      Education category
       Formal educationReferenceReferenceReferenceReference
       No formal education1.81 (1.45–2.26)< .0011.73 (1.23–2.43).0021.94 (1.27–2.96).0021.71 (1.11–2.64).015
      BMI = body mass index; CI = confidence interval; OR= odds ratio.
      a Inclusive of bilateral visual impairment and bilateral blindness, based on the better-seeing eye.
      b In this model, ethnicity was also adjusted, in addition to covariates listed in the table.
      c Systemic comorbidities were classified based on the concurrent presence of 2 or more systemic conditions, namely, diabetes, hyperlipidemia, hypertension, chronic kidney disease, or cardiovascular disease. Evaluation of systemic comorbidities as exposure variable was performed in a separate model, where individual exposures of diabetes, hypertension, hyperlipidemia, chronic kidney disease, and cardiovascular disease were not included in the model.
      d Cognitive impairment was assessed in a separate model, including only subjects aged 60 years and older, while adjusting for the same set of covariates as original main model.
      Furthermore, among the previously mentioned significant factors, old age (≥60 years) had the highest PAR for presenting bilateral visual loss in the overall sample (28.9%), followed by lower income level (21.0%), lower education level (12.4%), female sex (11.9%), and cognitive impairment (7.5%) (Table 8). Comparing the PAR profile across ethnicities (Figure 7), the PAR of old age for presenting bilateral visual loss was higher in Malay (31.6%) and Chinese (34.4%) compared with Indian individuals (21.3%); PAR of female sex was substantially higher in Malay (19.0%) than in Indian (8.5%) and Chinese individuals (8.4%); PAR for cognitive impairment was substantially higher in Malay (11.7%) than in Chinese individuals (2.7%). The profile of PAR was largely similar for income, and education level throughout the 3 ethnic groups.
      Table 8Population-Attributable Risk for Risk Factors Associated With Presenting and Best-corrected Bilateral Visual Loss (US Definition).
      PAR for Presenting Bilateral Visual LossPAR for Best-Corrected Bilateral Visual Loss
      OverallMalayIndianChineseOverallMalayIndianChinese
      Old age (60 and older)28.931.621.334.465.171.646.376.2
      Ethnicity
       ChineseReference
       Malay17.6
       Indian4.0
      Female11.919.08.58.422.327.027.99.1
      Medical conditions
       Diabetes mellitus11.39.2
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      17.010.0
       Chronic kidney disease2.40.2
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      3.81.6
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      7.86.4
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      9.32.4
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
       Cognitive impairment
      Cognitive impairment was assessed in a separate model, including only subjects aged 60 years and older.
      7.511.77.32.714.413.319.811.0
      Housing type
       1–2 room public housing flat3.85.22.02.24.65.1
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      0.5
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      3.1
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      Monthly income category
       < S$200021.024.1
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      17.124.658.168.5
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      58.955.8
      Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      Education category
       No formal education12.412.510.411.123.222.820.721.7
      PAR = population-attributable risk.
      Only factors that were significant in original multiple regression model (Tables 6 and 7) were included for PAR analysis. Best-corrected bilateral visual loss inclusive of bilateral visual impairment and bilateral blindness, based on the better-seeing eye. All values are percentages (%).
      a Denotes not statistically significant in original ethnic-specific multiple logistic regression model.
      b Cognitive impairment was assessed in a separate model, including only subjects aged 60 years and older.
      Figure thumbnail gr7
      Figure 7Population-attributable risk for risk factors associated with presenting bilateral visual loss (visual impairment or blindness).
      Similarly, when evaluating best-corrected bilateral visual loss (Table 8), old age had the highest PAR (65.1%), followed by lower income level (58.1%), lower education level (23.2%), female sex (22.3%), cognitive impairment (14.4%), and diabetes (11.3%). The PAR profiles of old age, female sex, cognitive impairment, and diabetes differ with ethnicity (Figure 8). Old age yielded higher PAR in Malay (71.6%) and Chinese (76.2%), compared with Indian individuals (46.3%). The PAR of female sex was higher in Malay (27.0%) and Indian individuals (27.9%), compared with Chinese individuals (9.1%). The PARs of diabetes (17.0%) and cognitive impairment (19.8%) were both highest in Indian compared with Malay and Chinese individuals.
      Figure thumbnail gr8
      Figure 8Population-attributable risk for risk factors associated with best-corrected bilateral visual loss (visual impairment or blindness).
      In Table 9, we further assessed the joint effects of risk factors associated with best-corrected visual loss (while adjusted for the same set of covariates as Table 7). Significant interaction was observed between older age group (≥60 years) and specific systemic diseases on best-corrected bilateral VI or blindness (all P for interaction ≤ .003). Compared with individuals aged <60 years without diabetes, individuals aged ≥60 years with diabetes were 12.7 times (95% CI, 8.39–19.23) as likely to have best-corrected bilateral VI or blindness. On the other hand, older individuals with CKD were 14.13 times (95% CI, 9.85–20.26) more likely to have best-corrected bilateral VI or blindness, compared with younger individuals without CKD. In addition, compared with individuals aged <60 years without any systemic diseases, older individuals with 2 systemic comorbidities were 14.76 times (95% CI, 6.39–34.06) likely to have best-corrected bilateral visual loss; older individuals with 3 or more systemic comorbidities were 26.56 times (95% CI, 11.62–60.71) as likely to have best-corrected bilateral visual loss. Furthermore, we also observed significant interaction between diabetes and CKD on best-corrected bilateral visual loss (P interaction = .005). Compared with individuals without diabetes and CKD, diabetic individuals with concurrent CKD were 2.19 times as likely to have best-corrected bilateral visual loss (Supplementary Table 4). There were no significant interactions between demographic, systemic, or socioeconomic factors on presenting bilateral VI or blindness.
      Table 9Stratum-specific Odds Ratio (95% Confidence Interval) of Best-corrected Bilateral Visual Loss (US Definition) in Categorical Groups of Age and Systemic Conditions
      Age Group, y
      <60≥60
      Diabetes status
       Absent1.00 (reference)10.04 (6.76–14.92)
      Denotes P < .001.
       Present2.96 (1.76–5.01)
      Denotes P < .001.
      12.70 (8.39–19.23)
      Denotes P < .001.
      P interaction = .003
      Chronic kidney disease status
       Absent1.00 (reference)8.37 (6.04–11.58)
      Denotes P < .001.
       Present4.83 (2.47–9.42)
      Denotes P < .001.
      14.13 (9.85–20.26)
      Denotes P < .001.
      P interaction = .003
      Systemic comorbidities
      Systemic comorbidities were classified based on the concurrent presence of 2 or more systemic conditions, namely, diabetes, hyperlipidemia, hypertension, chronic kidney disease, or cardiovascular disease. Evaluation of systemic comorbidities as exposure variable was performed in a separate model, where individual exposures of diabetes, hypertension, hyperlipidemia, chronic kidney disease and cardiovascular disease were not included in the model.
       No systemic disease1.00 (reference)12.58 (4.92–32.16)
      Denotes P < .001.
       Any 1 systemic disease2.28 (0.88–5.90)16.79 (7.24 38.92)
      Denotes P < .001.
       Any 2 systemic diseases3.71 (1.43–9.65)
      Denotes P < .05.
      14.76 (6.39–34.06)
      Denotes P < .001.
       ≥ 3 systemic diseases9.51 (3.84–23.60)
      Denotes P < .001.
      26.56 (11.62–60.71)
      Denotes P < .001.
      P interaction = .006
      Outcome variable was inclusive of bilateral visual impairment and bilateral blindness, based on better-seeing eye. Analysis was adjusted for age, sex, ethnicity, hypertension, cardiovascular disease, hyperlipidemia, deafness, body mass index, smoking status, alcohol consumption, housing type, living alone status, monthly income, and education.
      a Denotes P < .001.
      b Systemic comorbidities were classified based on the concurrent presence of 2 or more systemic conditions, namely, diabetes, hyperlipidemia, hypertension, chronic kidney disease, or cardiovascular disease. Evaluation of systemic comorbidities as exposure variable was performed in a separate model, where individual exposures of diabetes, hypertension, hyperlipidemia, chronic kidney disease and cardiovascular disease were not included in the model.
      c Denotes P < .05.

      Discussion

      In this thesis, we evaluated the prevalence and causes of VI and blindness among adults aged ≥40 years in a multiethnic Asian population of Malay, Indian, and Chinese individuals living in Singapore. We showed that 18.4% and 0.9% had presenting bilateral VI and blindness, respectively; whereas 4.2% and 0.3% had best-corrected bilateral VI and blindness, respectively. Overall, Malay individuals had higher prevalence of bilateral VI and blindness compared with Indian and Chinese individuals. Prevalence of VI and blindness in SEED was slightly lower compared with most previous studies conducted in Asia, which mainly focused on rural regions. Uncorrected refractive error and cataract were the top causes for presenting VI and blindness; cataract and DR were the top causes for best-corrected VI and blindness; these were consistently observed across the 3 ethnic groups.
      Although some previous studies provided preliminary information on risk factors associated with VI, findings from this thesis are the most comprehensive to date in this aspect, covering a wide range of relevant age-related systemic, demographic, and socioeconomic factors. Older age, female sex, lower socioeconomic status, diabetes, CKD, cognitive impairment, and systemic comorbidities were associated with higher risk of visual loss. Moreover, for the first time, we provided new findings on joint risk factor profiling, and identified that old age when combined with systemic comorbidities was associated with multifold risk of visual loss. This information are extremely useful to better risk-stratify high-risk subgroups and to aid in prioritization of health resource allocation, especially for the aging population. Last, we determined the PAR estimates of risk factors for visual loss, and identified that lower income and education had the highest modifiable attributable risks for visual loss. Such information will greatly help to guide policy makers in effectively identifying interventional areas and target for VI and blindness.
      In totality, this thesis provides the first large population-based summary of trends and risk factors associated with VI and blindness in a multiethnic Asian population in Singapore, using multifaceted evaluations and analyses, and providing new insights on traits related to visual loss in Asians. These results will be useful and relevant in impacting the future planning and designing of eye health services and blindness prevention strategies for Asia's rapidly developing urban communities.

       Comparison With Other Asian and Western Studies

      In our study, Malay individuals had higher age-standardized prevalence rates than Indian and Chinese individuals for presenting bilateral VI, blindness, and best-corrected bilateral VI. There were higher proportions of Malay individuals who had lower education and income and were residing in smaller public housing flats compared with Indian and Chinese individuals. Taken together, this suggests that more eye health awareness enhancement and social interventions need to be targeted on this group of individuals. Furthermore, a previous Singapore study reported less active health-seeking behavior (indicated by lower participation rate for health screenings) in Malay compared with Indian and Chinese individuals,
      • Lim R.B.
      • Zheng H.
      • Yang Q.
      • Cook A.R.
      • Chia K.S.
      • Lim W.Y.
      Ethnic and gender specific life expectancies of the Singapore population, 1965 to 2009 - converging, or diverging?.
      further suggesting that more “upstream interventions” need to be introduced and administered for Malay residents in Singapore.
      Compared with other Asian studies conducted in the past 2 decades (Supplementary Table 1), prevalence rates of VI and blindness in SEED were generally lower. Specifically, when comparing the Singapore Chinese Eye Study with other Chinese studies, such as the Handan Eye Study and the China Nine Province Survey,
      • Huang S.
      • Zheng Y.
      • Foster P.J.
      • Huang W.
      • He M.
      Prevalence and causes of visual impairment in Chinese adults in urban southern China.
      • Zhao J.
      • Ellwein L.B.
      • Cui H.
      • et al.
      Prevalence of vision impairment in older adults in rural China: the China Nine-Province Survey.
      • Liang Y.B.
      • Friedman D.S.
      • Wong T.Y.
      • et al.
      Prevalence and causes of low vision and blindness in a rural chinese adult population: the Handan Eye Study.
      • Li J.
      • Zhong H.
      • Cai N.
      • et al.
      The prevalence and causes of visual impairment in an elderly Chinese Bai ethnic rural population: the Yunnan minority eye study.
      • Tang Y.
      • Wang X.
      • Wang J.
      • et al.
      Prevalence and causes of visual impairment in a Chinese adult population: the Taizhou Eye Study.
      • Zhao J.
      • Xu X.
      • Ellwein L.B.
      • et al.
      Prevalence of vision impairment in older adults in rural China in 2014 and comparisons with the 2006 China nine-province survey.
      the prevalence of VI and blindness (for both presenting and best-corrected) was relatively lower for Singaporean Chinese. Similarly, Singaporean Indian individuals had lower rates of VI and blindness than other landmark population Indian studies, such as the Andhra Pradesh Eye Study
      • Dandona R.
      • Dandona L.
      • Srinivas M.
      • et al.
      Moderate visual impairment in India: the Andhra Pradesh Eye Disease Study.
      and the Central India Eye and Medical Study.
      • Nangia V.
      • Jonas J.B.
      • Gupta R.
      • Khare A.
      • Sinha A.
      Visual impairment and blindness in rural central India: the Central India Eye and Medical Study.
      Overall, these differences are expected given the differences in accessibility and utilization of eye care services between Singapore and the previously mentioned areas/countries. On the other hand, when further comparing with other population-based studies in Western developed countries, the prevalence rates in SEED were largely similar to those reported in the Blue Mountains Eye Study,
      • Attebo K.
      • Mitchell P.
      • Smith W.
      Visual acuity and the causes of visual loss in Australia. The Blue Mountains Eye Study.
      Beaver Dam Eye Study,
      • Klein R.
      • Klein B.E.
      • Linton K.L.
      • De Mets D.L.
      The Beaver Dam Eye Study: visual acuity.
      and the Los Angeles Latino Eye Study.
      • Varma R.
      • Ying-Lai M.
      • Klein R.
      • Azen S.P.
      Los Angeles Latino Eye Study Group
      Prevalence and risk indicators of visual impairment and blindness in Latinos: the Los Angeles Latino Eye Study.
      In contrast, the Rotterdam Eye Study seemingly had slightly higher rates of best-corrected blindness (1.4%) and VI (7.3%), compared with SEED.
      • Varma R.
      • Ying-Lai M.
      • Klein R.
      • Azen S.P.
      Los Angeles Latino Eye Study Group
      Prevalence and risk indicators of visual impairment and blindness in Latinos: the Los Angeles Latino Eye Study.
      Nevertheless, it also should be noted that accurate comparisons across studies are limited by the inherent variations in population characteristics, and time period between studies. Thus, these observed differences also should also be interpreted with caution.
      In terms of causes of visual loss, the overall leading causes for presenting VI in SEED were undercorrected refractive error, followed by cataract. Cataract was the top cause for presenting blindness. These were consistent across the 3 ethnic groups. Similarly, the Global Burden of Disease Study reported that undercorrected refractive error and cataract were the 2 common causes for presenting VI and blindness across countries in the high-income Asia Pacific region (developed countries), South Asia, East Asia, and Southeast Asia.
      • Flaxman S.R.
      • Bourne R.R.A.
      • Resnikoff S.
      • et al.
      Global causes of blindness and distance vision impairment 1990-2020: a systematic review and meta-analysis.
      Taken together, it appears that undercorrected refractive error and cataract are the universal causes for presenting visual loss in most Asian regions, even in urban communities. Given that these 2 conditions are readily preventable, this calls for further scale-up in eye care services and health strategies to improve eye care accessibility and enhance the adequacy of eye care services for affected individuals, particularly among elderly individuals.
      On the other hand, in SEED, cataract was the leading cause for best-corrected VI and blindness. Multiple Asian studies similarly reported cataract as the top cause for best-corrected blindness.
      • Xu L.
      • Wang Y.
      • Li Y.
      • et al.
      Causes of blindness and visual impairment in urban and rural areas in Beijing: the Beijing Eye Study.
      • Liang Y.B.
      • Friedman D.S.
      • Wong T.Y.
      • et al.
      Prevalence and causes of low vision and blindness in a rural chinese adult population: the Handan Eye Study.
      • Tang Y.
      • Wang X.
      • Wang J.
      • et al.
      Prevalence and causes of visual impairment in a Chinese adult population: the Taizhou Eye Study.
      • Nangia V.
      • Jonas J.B.
      • Gupta R.
      • Khare A.
      • Sinha A.
      Visual impairment and blindness in rural central India: the Central India Eye and Medical Study.
      • Iwase A.
      • Araie M.
      • Tomidokoro A.
      • Yamamoto T.
      • Shimizu H.
      • Kitazawa Y.
      Prevalence and causes of low vision and blindness in a Japanese adult population: the Tajimi Study.
      • Hsu W.M.
      • Cheng C.Y.
      • Liu J.H.
      • Tsai S.Y.
      • Chou P.
      Prevalence and causes of visual impairment in an elderly Chinese population in Taiwan: the Shihpai Eye Study.
      These collectively indicate that cataract is a universal cause for best-corrected VI and blindness for urban and rural regions. This further emphasizes the importance of continually improving accessibility to cataract surgery and optimal cataract surgery outcomes, which will in turn help to further reduce cataract-related VI and blindness.
      Followed by cataract, DR and myopic maculopathy were the second and third leading causes for best-corrected VI. The high prevalence of diabetes in SEED (29.5%) possibly explains the substantial number of DR-related VI cases. This again highlights the impact of diabetes on VI,
      • Ko F.
      • Vitale S.
      • Chou C.
      • Cotch M.
      • Saaddine J.
      • Friedman D.S.
      Prevalence of nonrefractive visual impairment in us adults and associated risk factors, 1999-2002 and 2005-2008.
      • Yau J.W.
      • Rogers S.L.
      • Kawasaki R.
      • et al.
      Global prevalence and major risk factors of diabetic retinopathy.
      further reinforcing the importance of greater awareness, compliance, and regular vision screening among diabetic individuals.
      • Sabanayagam C.
      • Yip W.
      • Ting D.S.
      • Tan G.
      • Wong T.Y.
      Ten emerging trends in the epidemiology of diabetic retinopathy.
      On the other hand, other Asian studies also reported myopic maculopathy as the leading cause for best-corrected VI after cataract.
      • Xu L.
      • Wang Y.
      • Li Y.
      • et al.
      Causes of blindness and visual impairment in urban and rural areas in Beijing: the Beijing Eye Study.
      • Hsu W.M.
      • Cheng C.Y.
      • Liu J.H.
      • Tsai S.Y.
      • Chou P.
      Prevalence and causes of visual impairment in an elderly Chinese population in Taiwan: the Shihpai Eye Study.
      This was not observed in other Western population studies,
      • Klein R.
      • Klein B.E.
      • Linton K.L.
      • De Mets D.L.
      The Beaver Dam Eye Study: visual acuity.
      • Attebo K.
      • Mitchell P.
      • Smith W.
      Visual acuity and the causes of visual loss in Australia. The Blue Mountains Eye Study.
      • Varma R.
      • Ying-Lai M.
      • Klein R.
      • Azen S.P.
      Los Angeles Latino Eye Study Group
      Prevalence and risk indicators of visual impairment and blindness in Latinos: the Los Angeles Latino Eye Study.
      further suggesting myopic maculopathy as an Asian-specific retinopathy.
      • Verkicharla P.K.
      • Ohno-Matsui K.
      • Saw S.M.
      Current and predicted demographics of high myopia and an update of its associated pathological changes.

       Associations Between Demographic and Socioeconomic Status With Visual Loss

      We found that older age and female sex were independently associated with increased risk of bilateral VI or blindness, for both presenting and best-corrected vision. This was similarly observed across the 3 ethnic groups. These findings also were consistent with that reported by previous studies in Asian and Western populations.
      • Xu L.
      • Wang Y.
      • Li Y.
      • et al.
      Causes of blindness and visual impairment in urban and rural areas in Beijing: the Beijing Eye Study.
      • Liang Y.B.
      • Friedman D.S.
      • Wong T.Y.
      • et al.
      Prevalence and causes of low vision and blindness in a rural chinese adult population: the Handan Eye Study.
      • Varma R.
      • Ying-Lai M.
      • Klein R.
      • Azen S.P.
      Los Angeles Latino Eye Study Group
      Prevalence and risk indicators of visual impairment and blindness in Latinos: the Los Angeles Latino Eye Study.
      • Ko F.
      • Vitale S.
      • Chou C.
      • Cotch M.
      • Saaddine J.
      • Friedman D.S.
      Prevalence of nonrefractive visual impairment in us adults and associated risk factors, 1999-2002 and 2005-2008.
      Notably, in our overall sample, old age (≥60 years) had the highest PAR for both presenting and best-corrected visual loss, further corroborating the significance of aging on the risk of visual loss.
      In the aspects of socioeconomic status, smaller housing, lower income, and lower educational level were associated with both presenting and best-corrected visual loss in our study. This observation was largely similar across the 3 ethnic groups in our sample. Likewise, previous studies also reported that lower education level and lower income were associated with VI and blindness.
      • Xu L.
      • Wang Y.
      • Li Y.
      • et al.
      Causes of blindness and visual impairment in urban and rural areas in Beijing: the Beijing Eye Study.
      • Ehrlich J.R.
      • Stagg B.C.
      • Andrews C.
      • Kumagai A.
      • Musch D.C.
      Vision impairment and receipt of eye care among older adults in low- and middle-income countries.
      • Ko F.
      • Vitale S.
      • Chou C.
      • Cotch M.
      • Saaddine J.
      • Friedman D.S.
      Prevalence of nonrefractive visual impairment in us adults and associated risk factors, 1999-2002 and 2005-2008.
      One possible explanation is that individuals with lower educational level may have limited awareness and lack of understanding of their own medical conditions, and thus less likely to go for routine vision screening, and have poorer compliance on follow-up even when referred to tertiary eye care for further treatment.
      • Newcomb P.A.
      • Klein R.
      Factors associated with compliance following diabetic eye screening.
      In addition, as reported previously, lower income is associated with poorer economic resources, which in turn may lead to poorer accessibility and utility of eye care services, subsequently resulting in suboptimal vision among economically disadvantaged groups in the long term.
      • Ko F.
      • Vitale S.
      • Chou C.
      • Cotch M.
      • Saaddine J.
      • Friedman D.S.
      Prevalence of nonrefractive visual impairment in us adults and associated risk factors, 1999-2002 and 2005-2008.
      • Taylor H.R.
      • Pezzullo M.L.
      • Keeffe J.E.
      The economic impact and cost of visual impairment in Australia.
      • Frick K.D.
      • Gower E.W.
      • Kempen J.H.
      • Wolff J.L.
      Economic impact of visual impairment and blindness in the United States.
      Furthermore, lower education and income also had been shown to be linked with less active health-seeking behavior,
      • Chan C.Q.H.
      • Lee K.H.
      • Low L.L.
      A systematic review of health status, health seeking behaviour and healthcare utilisation of low socioeconomic status populations in urban Singapore.
      which may in part explain lower utility of health services among the disadvantaged groups as well. More important, visual loss, when untreated, is likely to further hinder individuals' work productivity and earning potential, potentially making these individuals more vulnerable to visual loss even for preventable or treatable causes, such as undercorrected refractive error and cataract. Hence, without appropriate interventions, socioeconomically disadvantaged individuals, especially elderly individuals, may continue to be mired with challenges associated with VI, and thus have poorer quality of life.
      Of special note, our study in addition demonstrated that lower income and education were 2 of the highest modifiable attributable risk for presenting and best-corrected visual loss, further indicating that these 2 factors are suitable targets for public health intervention in reducing VI and blindness even for urban communities. In addition, when comparing with ethnicity, lower income and education still stood out to have higher PAR for best-corrected visual loss. This potentially suggests that the risk of best-corrected visual loss is more likely attributed to environmental factors (ie, socioeconomic related) than genetic ancestral factors.

       Systemic Diseases and Visual Loss

      In the overall sample of our study, we observed that individuals with diabetes were 1.4 times as likely to have best-corrected bilateral visual loss. Consistent with this, we also observed that the prevalence rates of bilateral VI and blindness for both presenting and best-corrected vision were higher in diabetic individuals compared with nondiabetic individuals in SEED (all P < .001) (Supplementary Table 5). The association between diabetes and best-corrected visual loss was also consistently observed across Malay, Indian, and Chinese individuals. Previous studies similarly reported higher prevalence of VI among diabetic individuals, and highlighted diabetes as a significant risk factor for nonrefractive VI.
      • Ko F.
      • Vitale S.
      • Chou C.
      • Cotch M.
      • Saaddine J.
      • Friedman D.S.
      Prevalence of nonrefractive visual impairment in us adults and associated risk factors, 1999-2002 and 2005-2008.
      • Kyari F.
      • Tafida A.
      • Sivasubramaniam S.
      • Murthy G.V.
      • Peto T.
      • Gilbert C.E.
      Prevalence and risk factors for diabetes and diabetic retinopathy: results from the Nigeria national blindness and visual impairment survey.
      • Zhang X.
      • Gregg E.W.
      • Cheng Y.J.
      • et al.
      Diabetes mellitus and visual impairment: National health and nutrition examination survey, 1999-2004.
      • Cui Y.
      • Zhang L.
      • Zhang M.
      • et al.
      Prevalence and causes of low vision and blindness in a Chinese population with type 2 diabetes: the Dongguan Eye Study.
      These observations in diabetes may be explained by the links between diabetes and several common age-related eye diseases, notably DR,
      • Lian J.X.
      • Gangwani R.A.
      • McGhee S.M.
      • Chan C.K.
      • Lam C.L.
      • Wong D.S.
      Systematic screening for diabetic retinopathy (DR) in Hong Kong: prevalence of DR and visual impairment among diabetic population.
      • Klein R.
      • Lee K.E.
      • Gangnon R.E.
      • Klein B.E.
      The 25-year incidence of visual impairment in type 1 diabetes mellitus the wisconsin epidemiologic study of diabetic retinopathy.
      cataract,
      • Asbell P.A.
      • Dualan I.
      • Mindel J.
      • Brocks D.
      • Ahmad M.
      • Epstein S.
      Age-related cataract.
      AMD,
      • Topouzis F.
      • Anastasopoulos E.
      • Augood C.
      • et al.
      Association of diabetes with age-related macular degeneration in the EUREYE study.
      and glaucoma.
      • Rim T.H.
      • Lee S.Y.
      • Bae H.W.
      • Seong G.J.
      • Kim S.S.
      • Kim C.Y.
      Increased risk of open-angle glaucoma among patients with diabetes mellitus: a 10-year follow-up nationwide cohort study.
      • Zhao D.
      • Cho J.
      • Kim M.H.
      • Friedman D.S.
      • Guallar E.
      Diabetes, fasting glucose, and the risk of glaucoma: a meta-analysis.
      In fact, DR was noted as the second leading cause of best-corrected VI in our study, consistent across the 3 ethnic groups. Furthermore, 11.3% of risk of best-corrected visual loss was attributed to diabetes in our sample, with an even higher PAR among Indian individuals (17.0%). Taken together, given diabetes is a modifiable risk factor, this further reinforces the importance of early identification of prediabetes, prevention of diabetes, and optimal diabetes management. Current efforts in regular vision screening and retinal photo examination
      • Nguyen H.V.
      • Tan G.S.
      • Tapp R.J.
      • et al.
      Cost-effectiveness of a national telemedicine diabetic retinopathy screening program in Singapore.
      among diabetic individuals should continue and increase, in a bid to detect VI and its underlying cause(s) at an early stage, allowing earlier interventions to be administered. This is especially important considering the growing prevalence of diabetes in urban communities.
      • Guariguata L.
      • Whiting D.R.
      • Hambleton I.
      • Beagley J.
      • Linnenkamp U.
      • Shaw J.E.
      Global estimates of diabetes prevalence for 2013 and projections for 2035.
      On the other hand, we also observed that CKD was associated with increased risk of both presenting and best-corrected bilateral visual loss. However, this association was observed only in Indian individuals (OR, 1.75). Similarly, in the Beijing Eye Study, Jonas and colleagues
      • Jonas J.B.
      • Wang Y.X.
      • Wei W.B.
      • Xu J.
      • You Q.S.
      • Xu L.
      Chronic kidney disease and eye diseases: the Beijing Eye Study.
      did not find significant association between CKD and VA in Chinese individuals. In the face of limited literature in this aspect currently, our findings provide preliminary indication that CKD is an independent risk factor for VI, a trait that may be unique to Indian individuals. Nevertheless, further studies, particularly longitudinal cohort, are warranted to better elucidate this.
      In addition, we showed that systemic comorbidities (concurrent presence of any 3 conditions of diabetes, hypertension, hyperlipidemia, CKD, and CVD) were associated with higher risk of best-corrected bilateral visual loss. This was, again, particularly marked among Indian individuals (OR, 4.03). On this front, we further observed significant interaction effects between diabetes and CKD. Individuals with concomitant diabetes and CKD were 2.2 times as likely to have best-corrected bilateral visual loss, compared with individuals free of both diseases. One possible explanation is that diabetes and CKD are closely linked with common pathophysiology, particularly the mechanisms in microvascular complications.
      • Wong C.W.
      • Wong T.Y.
      • Cheng C.Y.
      • Sabanayagam C.
      Kidney and eye diseases: common risk factors, etiological mechanisms, and pathways.
      • Yip W.
      • Sabanayagam C.
      • Ong P.G.
      • et al.
      Joint effect of early microvascular damage in the eye & kidney on risk of cardiovascular events.
      Similar to diabetes, CKD was also previously reported to be associated with age-related eye diseases, such as DR, AMD, and glaucoma,
      • Wong C.W.
      • Wong T.Y.
      • Cheng C.Y.
      • Sabanayagam C.
      Kidney and eye diseases: common risk factors, etiological mechanisms, and pathways.
      which are common causes of visual loss.

       Cognitive Impairment With Visual Loss

      In addition, we observed that individuals with cognitive impairment were 2.1 times as likely to have both presenting and best-corrected bilateral visual loss. This finding was similarly observed by Jonas and colleagues
      • Jonas J.B.
      • Wei W.B.
      • Zhu L.P.
      • Xu L.
      • Wang Y.X.
      Cognitive function and ophthalmological diseases: the Beijing Eye Study.
      in the Beijing Eye Study, which reported that subjects with decreased cognitive function scores had poorer presenting VA and BCVA. Previous studies also indicated that age-related ocular diseases and features, such as retinal degeneration, tessellated fundus, and suboptimal electrophysiological responses, were correlated with cognitive decline.
      • Jonas J.B.
      • Wei W.B.
      • Zhu L.P.
      • Xu L.
      • Wang Y.X.
      Cognitive function and ophthalmological diseases: the Beijing Eye Study.
      • Stothart G.
      • Kazanina N.
      • Naatanen R.
      • Haworth J.
      • Tales A.
      Early visual evoked potentials and mismatch negativity in Alzheimer's disease and mild cognitive impairment.
      In addition, severe forms of cognitive impairment, such as Alzheimer disease, had also been found to be associated with decreased vision, further corroborating a link between cognitive impairment and visual loss.
      • Fernandez R.
      • Kavcic V.
      • Duffy C.J.
      Neurophysiologic analyses of low- and high-level visual processing in Alzheimer disease.
      On the other hand, other prospective studies also reported that VI was associated with higher risk of incident cognitive impairment.
      • Ong S.Y.
      • Cheung C.Y.
      • Li X.
      • et al.
      Visual impairment, age-related eye diseases, and cognitive function: the Singapore Malay Eye study.
      • Chen S.P.
      • Bhattacharya J.
      • Pershing S.
      Association of vision loss with cognition in older adults.
      • Zheng D.D.
      • Swenor B.K.
      • Christ S.L.
      • West S.K.
      • Lam B.L.
      • Lee D.J.
      Longitudinal associations between visual impairment and cognitive functioning: the Salisbury Eye Evaluation Study.
      • Mitoku K.
      • Masaki N.
      • Ogata Y.
      • Okamoto K.
      Vision and hearing impairments, cognitive impairment and mortality among long-term care recipients: a population-based cohort study.
      In this regard, due to the cross-sectional nature of our study, we are unable to determine the actual causal relationship between cognitive impairment and VI. This warrants further longitudinal studies.

       Joint Effects of Age and Systemic Diseases on Visual Loss

      To the best of our knowledge, this is the first study that demonstrated significant interaction and effect modification between aging and systemic diseases on the risk of visual loss. Old age, combined with diabetes, CKD, or systemic comorbidities, was shown to have multifold risk of having best-corrected bilateral visual loss. This finding is especially relevant to the aging population in Asia and worldwide, particularly so for those residing in urban communities. Furthermore, this information may potentially aid in the formulation of more targeted screening strategies for high-risk subgroups. Specifically, individuals aged ≥60 years with concurrent diabetes, CKD, or any 2 systemic comorbidities (in the form of diabetes, CKD, hypertension, CVD, or hyperlipidemia) should be warranted for even more regular vision screenings to identify VI or blindness at an earlier stage. In addition, these findings also may help to guide decisions on better prioritization of screening or intervention strategies, when faced with limited resources.

       Strength and Limitations

      This thesis covers the 3 major ethnic groups in Asia (Malay, Indian, and Chinese), while adopting well-established and standardized examination methods and protocols across the 3 examined groups. This provides a unique and unprecedented opportunity to pool and compare data across the 3 ethnic groups in a relatively common geographic and socioeconomic environment. Other strengths of this thesis include large population-based samples that provide sufficient statistical power to comprehensively evaluate the independent and combined effects of multiple comorbidity measures of ethnicity, demographic, systemic, and socioeconomic risk factors on VI and blindness. Furthermore, the wide range of data collected in this study allowed for relevant confounding factors to be accounted for in the analysis, further substantiating the validity of our findings. Last, definition of VI and blindness (outcomes of interest) and diagnosis of eye diseases were defined according to international recommended criteria and guidelines, limiting the potential discrepancy and bias due to misclassification in our study.
      Nevertheless, this thesis also has a few limitations. First, this study was cross-sectional in nature, thus causal inference may not be conclusively drawn from current findings. Second, our analysis focused only on Asian ethnicity; replication of our findings in other populations may be required to further test for generalization. Third, as automated perimetry was performed on only a subset of study participants identified to be glaucoma suspects, loss of visual fields was not included as part of the definition for blindness. Therefore, the prevalence of blindness due to glaucoma may have been underestimated. Fourth, as cataract accounted for most VI cases after refractive correction, it was possible that concurrent retinal diseases might have been masked by significant cataract, especially among elderly individuals. Last, as described in the Methods section, the nonparticipants were older than the attended participants, thus potential bias in this study sample may not be entirely ruled out.

       Future Directions

      Findings from this thesis, in particular ethnic differences and risk factors for VI and blindness, may serve as good references for other population research groups in Asia and elsewhere. In addition, the direct and indirect economic impact of VI and blindness on communities and countries should be further analyzed and quantified comprehensively, to further provide holistic perspectives on the burden of visual loss. Such reports are currently well published for Western populations but still lacking for Asian populations. Further collaborations between research groups to combine data and to perform meta-analysis on VI remain necessary to advance the research and the understanding on the magnitude of VI in Asia, in particular urban communities where data remain scarce.

      Conclusion

      Asia is home to the world's largest population and accounts for the greatest burden of visual loss globally. In this multiethnic Asian population-based study in Singapore, we determined shared and unique traits associated with visual loss across Malay, Indian, and Chinese individuals. We further identified combined effects of age and systemic comorbidities on VI and blindness; and socioeconomic factors that had high modifiable attributable risks for visual loss. Collectively, these findings will be useful for the planning of eye health services, and designing of relevant interventions for Asia's rapidly developing urban communities, in a bid to reduce the burden of VI and blindness.
      All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. Funding/Support: The study is funded by National Medical Research Council (grants 0796/2003 , IRG07nov013 , IRG09nov014 , STaR/0003/2008 ; CG/SERI/2010 ) and Biomedical Research Council (grants 08/1/35/19/550 , 09/1/35/19/616 ), Singapore. The sponsor or funding organization had no role in the design or conduct of this research. All authors attest that they meet the current ICMJE requirements to qualify as authors.
      The authors thank Ms Miao-Li Chee (Singapore Eye Research Institute) for statistical consultation and support.

      Supplemental Data

      References

        • Bourne R.R.A.
        • Flaxman S.R.
        • Braithwaite T.
        • et al.
        Magnitude, temporal trends, and projections of the global prevalence of blindness and distance and near vision impairment: a systematic review and meta-analysis.
        Lancet Glob Health. 2017; 5: e888-e897
      1. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: a systematic analysis for the Global Burden of Disease Study 2015.
        Lancet. 2016; 388: 1545-1602
        • Lamoureux E.L.
        • Chong E.W.
        • Thumboo J.
        • et al.
        Vision impairment, ocular conditions, and vision-specific function: the Singapore Malay Eye Study.
        Ophthalmology. 2008; 115: 1973-1981
        • Siantar R.G.
        • Cheng C.Y.
        • Gemmy Cheung C.M.
        • et al.
        Impact of visual impairment and eye diseases on mortality: the Singapore Malay Eye Study (SiMES).
        Sci Rep. 2015; 5: 16304
        • Dai W.
        • Tham Y.C.
        • Chee M.L.
        • et al.
        Falls and recurrent falls among adults in a multi-ethnic Asian population: The Singapore Epidemiology of Eye Diseases Study.
        Sci Rep. 2018; 8: 7575
        • Pascolini D.
        • Mariotti S.P.
        Global estimates of visual impairment: 2010.
        Br J Ophthalmol. 2012; 96: 614-618
        • Wong W.L.
        • Su X.
        • Li X.
        • et al.
        Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis.
        Lancet Glob Health. 2014; 2: e106-e116
        • Tham Y.C.
        • Li X.
        • Wong T.Y.
        • Quigley H.A.
        • Aung T.
        • Cheng C.Y.
        Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis.
        Ophthalmology. 2014; 121: 2081-2090
        • United Nations Department of Economic and Social Affairs
        World Population Prospects: The 2017 Revision.
        United Nations, New York, NY2017
        • Bureau P.R.
        World Population Data Sheet.
        2013 (Available at https://www.prb.org/2013-world-population-data-sheet/. Accessed July 10, 2019)
        • Rutzen A.R.
        • Ellish N.J.
        • Schwab L.
        • et al.
        Blindness and eye disease in Cambodia.
        Ophthalmic Epidemiol. 2007; 14: 360-366
        • Zhao J.
        • Jia L.
        • Sui R.
        • Ellwein L.B.
        Prevalence of blindness and cataract surgery in Shunyi County, China.
        Am J Ophthalmol. 1998; 126: 506-514
        • Xu L.
        • Wang Y.
        • Li Y.
        • et al.
        Causes of blindness and visual impairment in urban and rural areas in Beijing: the Beijing Eye Study.
        Ophthalmology. 2006; 113: 1134.e1-11
        • Huang S.
        • Zheng Y.
        • Foster P.J.
        • Huang W.
        • He M.
        Prevalence and causes of visual impairment in Chinese adults in urban southern China.
        Arch Ophthalmol. 2009; 127: 1362-1367
        • Zhao J.
        • Ellwein L.B.
        • Cui H.
        • et al.
        Prevalence of vision impairment in older adults in rural China: the China Nine-Province Survey.
        Ophthalmology. 2010; 117 (416.e1): 409-416
        • Liang Y.B.
        • Friedman D.S.
        • Wong T.Y.
        • et al.
        Prevalence and causes of low vision and blindness in a rural chinese adult population: the Handan Eye Study.
        Ophthalmology. 2008; 115: 1965-1972
        • Li Z.
        • Cui H.
        • Liu P.
        • Zhang L.
        • Yang H.
        • Zhang L.
        Prevalence and causes of blindness and visual impairment among the elderly in rural southern Harbin, China.
        Ophthalmic Epidemiol. 2008; 15: 334-338
        • Li X.
        • Zhou Q.
        • Sun L.
        • et al.
        Prevalence of blindness and low vision in a rural population in northern China: preliminary results from a population-based survey.
        Ophthalmic Epidemiol. 2012; 19: 272-277
        • Li J.
        • Zhong H.
        • Cai N.
        • et al.
        The prevalence and causes of visual impairment in an elderly Chinese Bai ethnic rural population: the Yunnan minority eye study.
        Invest Ophthalmol Vis Sci. 2012; 53: 4498-4504
        • Tang Y.
        • Wang X.
        • Wang J.
        • et al.
        Prevalence and causes of visual impairment in a Chinese adult population: the Taizhou Eye Study.
        Ophthalmology. 2015; 122: 1480-1488
        • Yang W.Y.
        • Li J.
        • Zhao C.H.
        • et al.
        Population-based assessment of visual impairment among ethnic Dai adults in a rural community in China.
        Sci Rep. 2016; 6: 22590
        • Zhao J.
        • Xu X.
        • Ellwein L.B.
        • et al.
        Prevalence of vision impairment in older adults in rural China in 2014 and comparisons with the 2006 China nine-province survey.
        Am J Ophthalmol. 2018; 185: 81-93
        • Thulasiraj R.D.
        • Nirmalan P.K.
        • Ramakrishnan R.
        • et al.
        Blindness and vision impairment in a rural south Indian population: the Aravind Comprehensive Eye Survey.
        Ophthalmology. 2003; 110: 1491-1498
        • Murthy G.V.
        • Gupta S.
        • Ellwein L.B.
        • Munoz S.R.
        • Bachani D.
        • Dada V.K.
        A population-based eye survey of older adults in a rural district of Rajasthan: I. Central vision impairment, blindness, and cataract surgery.
        Ophthalmology. 2001; 108: 679-685
        • Thulasiraj R.D.
        • Rahamathulla R.
        • Saraswati A.
        • Selvaraj S.
        • Ellwein L.B.
        The Sivaganga eye survey: I. Blindness and cataract surgery.
        Ophthalmic Epidemiol. 2002; 9: 299-312
        • Dandona R.
        • Dandona L.
        • Srinivas M.
        • et al.
        Moderate visual impairment in India: the Andhra Pradesh Eye Disease Study.
        Br J Ophthalmol. 2002; 86: 373-377
        • Nirmalan P.K.
        • Thulasiraj R.D.
        • Maneksha V.
        • et al.
        A population based eye survey of older adults in Tirunelveli district of south India: blindness, cataract surgery, and visual outcomes.
        Br J Ophthalmol. 2002; 86: 505-512
        • Vijaya L.
        • George R.
        • Arvind H.
        • et al.
        Prevalence and causes of blindness in the rural population of the Chennai Glaucoma Study.
        Br J Ophthalmol. 2006; 90: 407-410
        • Nangia V.
        • Jonas J.B.
        • Gupta R.
        • Khare A.
        • Sinha A.
        Visual impairment and blindness in rural central India: the Central India Eye and Medical Study.
        Acta Ophthalmol. 2013; 91: 483-486
        • Iwase A.
        • Araie M.
        • Tomidokoro A.
        • Yamamoto T.
        • Shimizu H.
        • Kitazawa Y.
        Prevalence and causes of low vision and blindness in a Japanese adult population: the Tajimi Study.
        Ophthalmology. 2006; 113: 1354-1362
        • Nakamura Y.
        • Tomidokoro A.
        • Sawaguchi S.
        • Sakai H.
        • Iwase A.
        • Araie M.
        Prevalence and causes of low vision and blindness in a rural Southwest Island of Japan: the Kumejima study.
        Ophthalmology. 2010; 117: 2315-2321
        • Rim T.H.
        • Nam J.S.
        • Choi M.
        • Lee S.C.
        • Lee C.S.
        Prevalence and risk factors of visual impairment and blindness in Korea: the Fourth Korea National Health and Nutrition Examination Survey in 2008-2010.
        Acta Ophthalmol. 2014; 92: e317-e325
        • Thapa S.S.
        • Berg R.V.
        • Khanal S.
        • et al.
        Prevalence of visual impairment, cataract surgery and awareness of cataract and glaucoma in Bhaktapur district of Nepal: the Bhaktapur Glaucoma Study.
        BMC Ophthalmol. 2011; 11: 2
        • Saw S.M.
        • Foster P.J.
        • Gazzard G.
        • Seah S.
        Causes of blindness, low vision, and questionnaire-assessed poor visual function in Singaporean Chinese adults: The Tanjong Pagar Survey.
        Ophthalmology. 2004; 111: 1161-1168
        • Ahmad K.
        • Khan M.D.
        • Qureshi M.B.
        • et al.
        Prevalence and causes of blindness and low vision in a rural setting in Pakistan.
        Ophthalmic Epidemiol. 2005; 12: 19-23
        • Edussuriya K.
        • Sennanayake S.
        • Senaratne T.
        • et al.
        The prevalence and causes of visual impairment in central Sri Lanka the Kandy Eye study.
        Ophthalmology. 2009; 116: 52-56
        • Hsu W.M.
        • Cheng C.Y.
        • Liu J.H.
        • Tsai S.Y.
        • Chou P.
        Prevalence and causes of visual impairment in an elderly Chinese population in Taiwan: the Shihpai Eye Study.
        Ophthalmology. 2004; 111: 62-69
        • Anderson J.J.
        • Felson D.T.
        Factors associated with osteoarthritis of the knee in the first national Health and Nutrition Examination Survey (HANES I). Evidence for an association with overweight, race, and physical demands of work.
        Am J Epidemiol. 1988; 128: 179-189
        • Johnston J.
        Improving utilisation of dental services by understanding cultural difference.
        Int Dent J. 1993; 43: 506-511
        • Otten M.W.
        • Teutsch S.M.
        • Williamson D.F.
        • Marks J.S.
        The effect of known risk factors on the excess mortality of black adults in the United States.
        JAMA. 1990; 263: 845-850
        • Goff D.C.
        • Lloyd-Jones D.M.
        • Bennett G.
        • et al.
        2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines.
        J Am Coll Cardiol. 2014; 63: 2935-2959
      2. Edinburgh: NHS Health Scotland SG, 2009,1–84. The Scottish Ethnicity and Health Research Strategy Working Group. Health in our multi-ethnic Scotland future research priorities (full report). 2009. Available at http://www.healthscotland.com/documents/5720.aspx. Accessed July 10, 2019.

        • Tan C.E.
        • Emmanuel S.C.
        • Tan B.Y.
        • Jacob E.
        Prevalence of diabetes and ethnic differences in cardiovascular risk factors. The 1992 Singapore National Health Survey.
        Diabetes Care. 1999; 22: 241-247
        • Yeo K.
        • Tai B.
        • Heng D.
        • et al.
        Ethnicity modifies the association between diabetes mellitus and ischaemic heart disease in Chinese, Malays and Asian Indians living in Singapore.
        Diabetologia. 2006; 49: 2866-2873
        • Fisher D.E.
        • Shrager S.
        • Shea S.J.
        • et al.
        Visual impairment in White, Chinese, Black, and Hispanic participants from the Multi-Ethnic Study of Atherosclerosis Cohort.
        Ophthalmic Epidemiol. 2015; 22: 321-332
        • Tielsch J.M.
        • Sommer A.
        • Witt K.
        • Katz J.
        • Royall R.M.
        Blindness and visual impairment in an American urban population. The Baltimore Eye Survey.
        Arch Ophthalmol. 1990; 108: 286-290
        • Klein R.
        • Klein B.E.
        The prevalence of age-related eye diseases and visual impairment in aging: current estimates.
        Invest Ophthalmol Vis Sci. 2013; 54: ORSF5-ORSF13
        • Tham Y.C.
        • Liao J.
        • Vithana E.N.
        • et al.
        Aggregate effects of intraocular pressure and cup-to-disc ratio genetic variants on glaucoma in a multiethnic Asian Population.
        Ophthalmology. 2015; 122: 1149-1157
        • Wu D.
        • Dou J.
        • Chai X.
        • et al.
        Large-scale whole-genome sequencing of three diverse Asian populations in Singapore.
        bioRxiv. 2018; : 390070
        • Deng L.
        • Hoh B.P.
        • Lu D.
        • et al.
        Dissecting the genetic structure and admixture of four geographical Malay populations.
        Sci Rep. 2015; 5: 14375
        • Tham Y.C.
        • Aung T.
        • Fan Q.
        • et al.
        Joint Effects of Intraocular Pressure and Myopia on Risk of Primary Open-Angle Glaucoma: The Singapore Epidemiology of Eye Diseases Study.
        Sci Rep. 2016; 6: 19320
        • Chua J.
        • Tham Y.C.
        • Liao J.
        • et al.
        Ethnic differences of intraocular pressure and central corneal thickness: the Singapore Epidemiology of Eye Diseases study.
        Ophthalmology. 2014; 121: 2013-2022
        • Tan G.S.
        • Gan A.
        • Sabanayagam C.
        • et al.
        Ethnic differences in the prevalence and risk factors of diabetic retinopathy: the Singapore Epidemiology of Eye Diseases Study.
        Ophthalmology. 2018; 125: 529-536
        • Foster P.J.
        • Oen F.T.
        • Machin D.
        • et al.
        The prevalence of glaucoma in Chinese residents of Singapore: a cross-sectional population survey of the Tanjong Pagar district.
        Arch Ophthalmol. 2000; 118: 1105-1111
        • Shen S.Y.
        • Wong T.Y.
        • Foster P.J.
        • et al.
        The prevalence and types of glaucoma in malay people: the Singapore Malay eye study.
        Invest Ophthalmol Vis Sci. 2008; 49: 3846-3851
        • Narayanaswamy A.
        • Baskaran M.
        • Zheng Y.
        • et al.
        The prevalence and types of glaucoma in an urban Indian population: the Singapore Indian Eye Study.
        Invest Ophthalmol Vis Sci. 2013; 54: 4621-4627
        • Wong T.Y.
        • Tielsch J.M.
        • Schein O.D.
        Racial difference in the incidence of retinal detachment in Singapore.
        Arch Ophthalmol. 1999; 117: 379-383
        • Wong T.Y.
        • Foster P.J.
        • Seah S.K.
        • Chew P.T.
        Rates of hospital admissions for primary angle closure glaucoma among Chinese, Malays, and Indians in Singapore.
        Br J Ophthalmol. 2000; 84: 990-992
        • Wong T.Y.
        Cataract extraction rates among Chinese, Malays, and Indians in Singapore: a population-based analysis.
        Arch Ophthalmol. 2001; 119: 727-732
        • Gong P.
        • Liang S.
        • Carlton E.J.
        • et al.
        Urbanisation and health in China.
        Lancet. 2012; 379: 843-852
        • Strong K.
        • Mathers C.
        • Leeder S.
        • Beaglehole R.
        Preventing chronic diseases: how many lives can we save?.
        Lancet. 2005; 366: 1578-1582
        • Ehrlich J.R.
        • Stagg B.C.
        • Andrews C.
        • Kumagai A.
        • Musch D.C.
        Vision impairment and receipt of eye care among older adults in low- and middle-income countries.
        (JAMA Ophthalmol) (2018.11.21)
        • Klein R.
        • Klein B.E.
        • Linton K.L.
        • De Mets D.L.
        The Beaver Dam Eye Study: visual acuity.
        Ophthalmology. 1991; 98: 1310-1315
        • Attebo K.
        • Mitchell P.
        • Smith W.
        Visual acuity and the causes of visual loss in Australia. The Blue Mountains Eye Study.
        Ophthalmology. 1996; 103: 357-364
        • Varma R.
        • Ying-Lai M.
        • Klein R.
        • Azen S.P.
        • Los Angeles Latino Eye Study Group
        Prevalence and risk indicators of visual impairment and blindness in Latinos: the Los Angeles Latino Eye Study.
        Ophthalmology. 2004; 111: 1132-1140
        • Bourne R.R.
        • Stevens G.A.
        • White R.A.
        • et al.
        Causes of vision loss worldwide, 1990-2010: a systematic analysis.
        Lancet Glob Health. 2013; 1: e339-e349
        • Flaxman S.R.
        • Bourne R.R.A.
        • Resnikoff S.
        • et al.
        Global causes of blindness and distance vision impairment 1990-2020: a systematic review and meta-analysis.
        Lancet Glob Health. 2017; 5: e1221-e1234
        • Goh S.Y.
        • Ang S.B.
        • Bee Y.M.
        • et al.
        Ministry of Health clinical practice guidelines: diabetes mellitus.
        Singapore Med J. 2014; 55: 334-347
        • Sabanayagam C.
        • Khoo E.Y.
        • Lye W.K.
        • et al.
        Diagnosis of diabetes mellitus using HbA1c in Asians: relationship between HbA1c and retinopathy in a multiethnic Asian population.
        J Clin Endocrinol Metab. 2015; 100: 689-696
        • Wong T.Y.
        • Aiello L.P.
        • Ferris F.
        • et al.
        ICO Guidelines for Diabetic Eye Care (Updated 2017).
        International Council of Ophthalmology, San Francisco, CA2017
        • Nguyen H.V.
        • Tan G.S.
        • Tapp R.J.
        • et al.
        Cost-effectiveness of a national telemedicine diabetic retinopathy screening program in Singapore.
        Ophthalmology. 2016; 123: 2571-2580
        • Ansah J.P.
        • De Korne D.
        • Bayer S.
        • et al.
        Future requirements for and supply of ophthalmologists for an aging population in Singapore.
        Hum Resour Health. 2015; 13: 86
        • Zheng Y.
        • Lavanya R.
        • Wu R.
        • et al.
        Prevalence and causes of visual impairment and blindness in an urban Indian population: the Singapore Indian Eye Study.
        Ophthalmology. 2011; 118: 1798-1804
        • Tham Y.C.
        • Lim S.H.
        • Shi Y.
        • et al.
        Trends of visual impairment and blindness in the Singapore Chinese population over a decade.
        Sci Rep. 2018; 8: 12224
        • Wong T.Y.
        • Chong E.W.
        • Wong W.L.
        • et al.
        Prevalence and causes of low vision and blindness in an urban malay population: the Singapore Malay Eye Study.
        Arch Ophthalmol. 2008; 126: 1091-1099
        • Lim R.B.
        • Zheng H.
        • Yang Q.
        • Cook A.R.
        • Chia K.S.
        • Lim W.Y.
        Ethnic and gender specific life expectancies of the Singapore population, 1965 to 2009 - converging, or diverging?.
        BMC Public Health. 2013; 13: 1012
        • Chan C.Q.H.
        • Lee K.H.
        • Low L.L.
        A systematic review of health status, health seeking behaviour and healthcare utilisation of low socioeconomic status populations in urban Singapore.
        Int J Equity Health. 2018; 17: 39
        • World Health Organization
        Universal Eye Health: A Global Action Plan, 2014-2019.
        World Health Organization, Geneva, Switzerland2013
        • Foong A.W.
        • Saw S.M.
        • Loo J.L.
        • et al.
        Rationale and methodology for a population-based study of eye diseases in Malay people: The Singapore Malay eye study (SiMES).
        Ophthalmic Epidemiol. 2007; 14: 25-35
        • Lavanya R.
        • Jeganathan V.S.
        • Zheng Y.
        • et al.
        Methodology of the Singapore Indian Chinese Cohort (SICC) eye study: quantifying ethnic variations in the epidemiology of eye diseases in Asians.
        Ophthalmic Epidemiol. 2009; 16: 325-336
        • Foster P.J.
        • Buhrmann R.
        • Quigley H.A.
        • Johnson G.J.
        The definition and classification of glaucoma in prevalence surveys.
        Br J Ophthalmol. 2002; 86: 238-242
        • Klein R.
        • Davis M.D.
        • Magli Y.L.
        • Segal P.
        • Klein B.E.
        • Hubbard L.
        The Wisconsin age-related maculopathy grading system.
        Ophthalmology. 1991; 98: 1128-1134
      3. Grading diabetic retinopathy from stereoscopic color fundus photographs--an extension of the modified Airlie House classification. ETDRS report number 10. Early Treatment Diabetic Retinopathy Study Research Group.
        Ophthalmology. 1991; 98: 786-806
        • Chylack Jr., L.T.
        • Wolfe J.K.
        • Singer D.M.
        • et al.
        The Lens Opacities Classification System III. The Longitudinal Study of Cataract Study Group.
        Arch Ophthalmol. 1993; 111: 831-836
        • Ong S.Y.
        • Cheung C.Y.
        • Li X.
        • et al.
        Visual impairment, age-related eye diseases, and cognitive function: the Singapore Malay Eye study.
        Arch Ophthalmol. 2012; 130: 895-900
      4. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification.
        Am J Kidney Dis. 2002; 39: S1-S266
        • Levey A.S.
        • Stevens L.A.
        • Schmid C.H.
        • et al.
        A new equation to estimate glomerular filtration rate.
        Ann Intern Med. 2009; 150: 604-612
        • Bruzzi P.
        • Green S.B.
        • Byar D.P.
        • Brinton L.A.
        • Schairer C.
        Estimating the population attributable risk for multiple risk factors using case-control data.
        Am J Epidemiol. 1985; 122: 904-914
        • Ko F.
        • Vitale S.
        • Chou C.
        • Cotch M.
        • Saaddine J.
        • Friedman D.S.
        Prevalence of nonrefractive visual impairment in us adults and associated risk factors, 1999-2002 and 2005-2008.
        JAMA. 2012; 308: 2361-2368
        • Yau J.W.
        • Rogers S.L.
        • Kawasaki R.
        • et al.
        Global prevalence and major risk factors of diabetic retinopathy.
        Diabetes Care. 2012; 35: 556-564
        • Sabanayagam C.
        • Yip W.
        • Ting D.S.
        • Tan G.
        • Wong T.Y.
        Ten emerging trends in the epidemiology of diabetic retinopathy.
        Ophthalmic Epidemiol. 2016; 23: 209-222
        • Verkicharla P.K.
        • Ohno-Matsui K.
        • Saw S.M.
        Current and predicted demographics of high myopia and an update of its associated pathological changes.
        Ophthalmic Physiol Opt. 2015; 35: 465-475
        • Newcomb P.A.
        • Klein R.
        Factors associated with compliance following diabetic eye screening.
        J Diabet Complications. 1990; 4: 8-14
        • Taylor H.R.
        • Pezzullo M.L.
        • Keeffe J.E.
        The economic impact and cost of visual impairment in Australia.
        Br J Ophthalmol. 2006; 90: 272-275
        • Frick K.D.
        • Gower E.W.
        • Kempen J.H.
        • Wolff J.L.
        Economic impact of visual impairment and blindness in the United States.
        Arch Ophthalmol. 2007; 125: 544-550
        • Kyari F.
        • Tafida A.
        • Sivasubramaniam S.
        • Murthy G.V.
        • Peto T.
        • Gilbert C.E.
        Prevalence and risk factors for diabetes and diabetic retinopathy: results from the Nigeria national blindness and visual impairment survey.
        BMC Public Health. 2014; 14: 1299
        • Zhang X.
        • Gregg E.W.
        • Cheng Y.J.
        • et al.
        Diabetes mellitus and visual impairment: National health and nutrition examination survey, 1999-2004.
        Arch Ophthalmol. 2008; 126: 1421-1427
        • Cui Y.
        • Zhang L.
        • Zhang M.
        • et al.
        Prevalence and causes of low vision and blindness in a Chinese population with type 2 diabetes: the Dongguan Eye Study.
        Sci Rep. 2017; 7: 11195
        • Lian J.X.
        • Gangwani R.A.
        • McGhee S.M.
        • Chan C.K.
        • Lam C.L.
        • Wong D.S.
        Systematic screening for diabetic retinopathy (DR) in Hong Kong: prevalence of DR and visual impairment among diabetic population.
        Br J Ophthalmol. 2016; 100: 151-155
        • Klein R.
        • Lee K.E.
        • Gangnon R.E.
        • Klein B.E.
        The 25-year incidence of visual impairment in type 1 diabetes mellitus the wisconsin epidemiologic study of diabetic retinopathy.
        Ophthalmology. 2010; 117: 63-70
        • Asbell P.A.
        • Dualan I.
        • Mindel J.
        • Brocks D.
        • Ahmad M.
        • Epstein S.
        Age-related cataract.
        Lancet. 2005; 365: 599-609
        • Topouzis F.
        • Anastasopoulos E.
        • Augood C.
        • et al.
        Association of diabetes with age-related macular degeneration in the EUREYE study.
        Br J Ophthalmol. 2009; 93: 1037-1041
        • Rim T.H.
        • Lee S.Y.
        • Bae H.W.
        • Seong G.J.
        • Kim S.S.
        • Kim C.Y.
        Increased risk of open-angle glaucoma among patients with diabetes mellitus: a 10-year follow-up nationwide cohort study.
        Acta Ophthalmol. 2018; 96: e1025-e1030
        • Zhao D.
        • Cho J.
        • Kim M.H.
        • Friedman D.S.
        • Guallar E.
        Diabetes, fasting glucose, and the risk of glaucoma: a meta-analysis.
        Ophthalmology. 2015; 122: 72-78
        • Guariguata L.
        • Whiting D.R.
        • Hambleton I.
        • Beagley J.
        • Linnenkamp U.
        • Shaw J.E.
        Global estimates of diabetes prevalence for 2013 and projections for 2035.
        Diabetes Res Clin Pract. 2014; 103: 137-149
        • Jonas J.B.
        • Wang Y.X.
        • Wei W.B.
        • Xu J.
        • You Q.S.
        • Xu L.
        Chronic kidney disease and eye diseases: the Beijing Eye Study.
        Ophthalmology. 2017; 124: 1566-1569
        • Wong C.W.
        • Wong T.Y.
        • Cheng C.Y.
        • Sabanayagam C.
        Kidney and eye diseases: common risk factors, etiological mechanisms, and pathways.
        Kidney Int. 2014; 85: 1290-1302
        • Yip W.
        • Sabanayagam C.
        • Ong P.G.
        • et al.
        Joint effect of early microvascular damage in the eye & kidney on risk of cardiovascular events.
        Sci Rep. 2016; 6: 27442
        • Jonas J.B.
        • Wei W.B.
        • Zhu L.P.
        • Xu L.
        • Wang Y.X.
        Cognitive function and ophthalmological diseases: the Beijing Eye Study.
        Sci Rep. 2018; 8: 4816
        • Stothart G.
        • Kazanina N.
        • Naatanen R.
        • Haworth J.
        • Tales A.
        Early visual evoked potentials and mismatch negativity in Alzheimer's disease and mild cognitive impairment.
        J Alzheimers Dis. 2015; 44: 397-408
        • Fernandez R.
        • Kavcic V.
        • Duffy C.J.
        Neurophysiologic analyses of low- and high-level visual processing in Alzheimer disease.
        Neurology. 2007; 68: 2066-2076
        • Chen S.P.
        • Bhattacharya J.
        • Pershing S.
        Association of vision loss with cognition in older adults.
        JAMA Ophthalmol. 2017; 135: 963-970
        • Zheng D.D.
        • Swenor B.K.
        • Christ S.L.
        • West S.K.
        • Lam B.L.
        • Lee D.J.
        Longitudinal associations between visual impairment and cognitive functioning: the Salisbury Eye Evaluation Study.
        JAMA Ophthalmol. 2018; 136: 989-995
        • Mitoku K.
        • Masaki N.
        • Ogata Y.
        • Okamoto K.
        Vision and hearing impairments, cognitive impairment and mortality among long-term care recipients: a population-based cohort study.
        BMC Geriatr. 2016; 16: 112